Contractual crowding out

There is good evidence that the level of detail in legal contracts may crowd out moral intentions in relation to the agreements that the contracts represent. And despite the evidence seen in my prior blog on the importance of fairness to the law (see Fairness piercing the veil, and filling gaps in the law), it seems that there are increasing moves to squeeze fairness out from legal considerations, at least in the US legal system. Unsurprisingly perhaps, that is only to the law’s detriment, and to society’s.

Pennsylvania law professor Tess Wilkinson-Ryan found that, generally, people are averse to breaching contracts, and typically believe that those in breach should make higher pay-outs than efficiency might indicate was necessary. Typically, contract law expects the wronged party to mitigate their loss – for example seeking an alternative provider of a contracted-for widget rather than just ceasing all production and blaming the delinquent supplier. But Wilkinson-Ryan finds that most people think that the delinquent supplier should be held accountable for more than this mitigated loss.

She also finds, however, that if a contract includes a liquidated damages clause, most people’s response to any contractual failure is less morally activated, and tends to accept that the damages built into the contract are appropriate. This particular consequence of the sense of fairness means that breach of contract becomes more acceptable: “it seems reasonable to conclude that subjects’ moral qualms deter efficient breach, but that the presence of a liquidated-damages clause in a contract reduces those qualms and, in turn, encourages breach”.

At its core, the finding is that: “it is not clear that breaching a contract with a liquidated-damages clause comes firmly within the category of breaking a promise”. Wilkinson-Ryan contrasts this attitude with that of a contractual breach that requires a lawsuit or the appointment of an authoritative body to assess what is the fair recompense for that breach. At such points, considerations of fairness and morality come into play; not so much if the contract has detailed an agreed level of damages.

A somewhat similar finding emerged from a sociological study considering trust in markets where contracts are complete or not complete. In this case, the quality of the relevant goods was variable. Where the contract enforced a level of quality, trust did not develop in supply relationships. However, in contrast, where the quality was not subject to contract, trust and fair treatment became vital parts of supply chain relations as the clear alternative provider of a degree of quality control. The development of such trusting and fair relationships had important and positive knock-on consequences: concern for one’s own reputation and considering others’ reputations, and a greater sense of commitment to trading partners. There are benefits from leaving space in contracts for fairness and trust to develop.

The long-term positives from such a commitment are shown in a second similar study, where again the quality of goods was either subject to contract or not capable of being contracted for. Where contracts were complete and included details on quality, trading relationships were less durable: 90% lasted less than 3 periods. In contrast, when trust and fair treatment were needed to fill the gaps in the contracts, trading relationships generally lasted much longer.

This consistent finding that deploying fairness rather than detailed legal agreements brings clear benefits makes the recent tendency in the US legal system to squeeze out fairness altogether still more concerning. In my forthcoming book, I highlight the way in which the US law in the areas of antitrust (rules against monopolist behaviours and other abuses of competitive position) and bankruptcy has been interpreted such that fairness no longer plays a part. Formerly, fairness featured heavily in the legal approach to both issues. The consequences of this shift aren’t great.

A key study considers recent changes to the US antitrust regime that have essentially meant it abandons considerations of fairness to consumers and focuses solely on narrow interpretations of the concept of economic efficiency. It argues that these changes mean that the law has failed “to control the nefarious consequences of unrestrained market power”.

With regard to bankruptcy, the key legal precedent was a Delaware Supreme Court decision in 2007, North American Catholic Educational Programming v Gheewalla. This essentially removed the previous prioritisation given to equitable treatment of creditors – an expectation of fairness – and required instead a strict contractual approach to be applied. Since that abandonment of fair treatment of creditors by failing companies, battles have ensued between different finance houses over their carcases (part of so-called ‘bankruptcy hardball’). Though the US Chapter 11 bankruptcy process in theory gives heavy weight to the post-bankruptcy survival and prosperity of the stricken business, often this battling between finance houses leaves little value within the business to give it any chance of later prosperity.

According to Harvard Law School professor Jared Ellias and Brown Rudnick lawyer Robert Stark, the consequences are profound:

“The slow moving trains of justice here have broader consequences than denying justice to one particular plaintiff or another. It emboldens the entire private equity industry to extract excessive dividends from portfolio firms, knowing that it might take more than a decade to litigate the fraudulent transfer action, by which time every employee currently at the private equity firm will be gone.”

Contracts are being written to crowd out fairness; the law (at least in the US) is often being understood to squeeze out fairness and lean instead on narrow contractual or economic readings. That isn’t good for the legal system, and it seems that it isn’t good for society either.

I am happy to confirm as ever that the Sense of Fairness blog is a purely personal endeavour

Do Liquidated Damages Encourage Breach? A Psychological Experiment, Tess Wilkinson-Ryan. Michigan Law Review 108, 2010

The Emergence of Exchange Structures: An Experimental Study of Uncertainty, Commitment, and Trust, Peter Kollock. American Journal of Sociology, 100, No 2, 1994

Relational Contracts and the Nature of Market Interactions, Martin Brown, Armin Falk, Ernst Fehr. Econometrica 72, No 3, 2004

The antitrust F word: Fairness considerations in competition law, Sandra Marco Colino. Journal of Business Law, 2019

North American Catholic Educational Programming Foundation, Inc v Gheewalla, Delaware Supreme Court 930 A.2d 92 (Del. 2007)

Bankruptcy Hardball, Jared Ellias, Robert Stark. 108 California Law Review 745, 2020

Fairness piercing the veil, and filling gaps in the law

An extraordinary legal case was decided in the UK in April, and it has recently been determined that there can be no appeal from it. The court decision has led directly to the insolvency of a major property development group, and it illustrates the way in which fairness fills gaps in the law – indeed how fairness is a core element of how the law works.

The case is extraordinary because it deploys fairness effectively to overthrow a fundamental element of corporate law. The decision in effect enables the piercing of the corporate veil – the legal barrier that means the liabilities of a company stop at its doors and cannot be claimed from a parent, other group company, or shareholder. Limited liability, through this mechanism of the corporate veil, is generally regarded as a core element of capitalism, a protection of the shareholder providing capital, freeing them from the risk that they might lose more than their contribution to the company through their buying of shares. Because of this perceived centrality to the capitalist system, courts, and indeed legislators, are deeply wary of piercing the veil in the way this decision does.

However, there are certain policy aims that are so substantial that such effective veil piercing is seen as necessary. That is true of the policy underlying the law in the case in question, Crest Nicholson Regeneration v Ardmore Construction. It concerns Building Liability Orders, established under the Building Safety Act to ensure that appropriate contributions are made by those responsible for the remediation of buildings which since the terrible Grenfell Tower fire (see Power leads us astray: fairness lessons from Grenfell) are recognised as representing significant risks. Building Liability Orders (BLOs) allow the courts to extend the liability for remediation costs from the immediate company involved to associated companies where it is ‘just and equitable’ (what this blog would call fair) to do so.

Image from *KCTMO – Feeling the Heat!*, sadly prescient blog from Grenfell Action Group, March 14 2017

I deliberately say that this ‘in effect’ pierces the corporate veil, because it’s worth noting that there is some debate as to whether this is a full piercing, though to a non-specialist this may seem a debate on the head of a pin. As Christopher Veal of leading barristers’ chambers Pump Court explains:

“Whether this constitutes “piercing the corporate veil” in the traditional sense is contested: some courts and commentators treat BLOs as a distinct statutory remedy rather than classic veil-piercing, on the basis that the court is not disregarding the corporate structure but rather applying a legislative override. Others characterise the practical effect as functionally equivalent to piercing the veil. The distinction matters less in practice than the statutory test itself, which turns on whether it is just and equitable to extend liability.”

Even though there is this debate, the case makes clear that sometimes statute and policy considerations will take decisions to extend liability notwithstanding the perceived strong constraint of the corporate veil. This step was seen as particularly needed in the development sector, where there is a history of each individual development being built by a thinly capitalised subsidiary so as to protect the parent, and wider group, from exposures to any potential liability. The immediate subsidiary involved in the development, Ardmore Construction Ltd, had been put into administration on the day before the initial adjudication against it (the judge decided to accept the suggestion that this was coincidence), and the court also found that there had been a corporate restructuring with some intent to limit broader liability. The ruling means that this restructuring has failed: following the court’s decision to pierce the veil, the whole Ardmore Construction Group entered administration in mid-June, and other related companies have sought protection from creditors.

This application of fairness to close gaps in the law isn’t as infrequent as people might imagine. In many ways, it is a core element of the functioning of the common law, under which judges reach decisions that are not explicitly covered by legislation, applying broader judgement and higher principles to do so. Fairness is often one of those higher principles. Think about the famous 1930s case of the snail in the bottle of ginger beer, Donoghue v Stevenson. There would have been no need for the case to reach the House of Lords (then the UK’s highest court) if the drinker of the ginger beer in question had been its purchaser rather than her friend for whom she bought it. If the buyer had drunk it and suffered the consequences, contract law protections would have applied. In the absence of some form of contractual protection for the drinker, the House of Lords applied a more general duty of care to manufacturers and found that the Stevenson business could be seen to have negligently breached that duty of care. From this finding based in fairness has arisen the whole modern concept of the law of tort, whereby in certain constrained circumstances duties of care arise outside of contractual relationships.

In other situations, courts will imply fairness into contracts that on the face of it do not expect it. In her interesting book Fool Proof, law professor Tess Wilkinson-Ryan discusses the Massachusetts Supreme Court case Fortune v National Cash Register. Orville Fortune was employed at will and at a low salary, with his main pay opportunity through commission on sales – a portion paid at the point of sale and the rest on delivery and installation of the equipment. Fortune was sacked between landing a big sale and the point at which he would have received the second part of his commission from that sale. While on the face of the contract his employer could indeed do this, the courts implied a duty of good faith such that National Cash Regiester was not in fact permitted to exploit its worker in this way. Instead, it needed to treat Fortune fairly and allow him to take the fair benefit of the commission he had earned through his sale. The court stated bluntly:

“we are merely recognizing the general requirement in this Commonwealth that parties to contracts and commercial transactions must act in good faith toward one another. Good faith and fair dealing between parties are pervasive requirements in our law; it can be said fairly, that parties to contracts or commercial transactions are bound by this standard.”

Further, there is a whole section of the English legal system fully animated by fairness. It is not by chance that this is called ‘equity’, now formally combined with the rest of the legal system but with its origins in a mediaeval court system focused on delivering fairness. As Alastair Hudson states in the opening words of the seventh edition of his Understanding Equity & Trusts: “Equity is a means by which English law ensures fair outcomes in individual cases where the strict application of the common law or statute would otherwise generate injustice.” In practice, the concept of equity applies somewhat more narrowly than this assertion implies, but the intent is to identify the fair outcome when there is a lack of clarity in the common law or in statute. Equity lies at the heart of trust law, and so as I explore in my forthcoming book, the sense of fairness animates the fiduciary duties with which trustees are vested.

Hudson develops the thought further in his Great Debates in Equity and Trusts, recognising the need for the equity approach to soften the edge cases of common law and statute, so as to ensure that the legal system overall is “rounded”. He argues: “Clearly, there need to be rules … However, there will also be situations in which the ‘letter of the law’ will not necessarily produce an ideal response, or worse, a positively unfair response. There will be situations in which people need some flexibility in which to create new models which will work better for them personally or better in unanticipated situations. One future for the UK economy is likely to be in the creative field and in the new frontiers of technology. By definition, an innovative economy may need innovative legal models … to maximise its success. Therefore, non-rigid ways of thinking will be important.”

He concludes: “a legal system needs a synthesis of strict rules (of the sort typified by statute and by common law) and mechanisms for achieving fair outcomes when the strict rules will not achieve that”.

The extent to which fairness will require further substantive upendings of traditional legal understandings such as the Crest Nicholson Regeneration v Ardmore Construction piercing of the corporate veil remains to be seen. But it seems clear that fairness will continue to be needed to fill gaps in, and round off the edges of, the common law and legislation.

I am happy to confirm as ever that the Sense of Fairness blog is a purely personal endeavour.

Crest Nicholson Regeneration v Ardmore Construction [2026] EWHC 789

Crest Nicholson v Ardmore [2026] EWHC 789 (TCC): A Landmark Decision on Building Liability Orders, Christopher Veal. Pump Court Chambers, 22 May 2026

Donoghue v Stevenson [1932] AC 562

Fool Proof: How Fear of Playing the Sucker Shapes our Selves and the Social Order – and what we can do about it, Tess Wilkinson-Ryan. Harper Collins, 2023

Fortune v National Cash Register Co, 373 Mass 96 (1977)

Understanding Equity & Trusts, Alastair Hudson. Routledge, Seventh Edition 2022

Great Debates in Equity and Trusts, Alastair Hudson. Macmillan, 2014

Squid Game ‘fairness’ II

This blogpost follows Squid Game series (season) 3 and while it doesn’t contain specific spoilers it does assume some knowledge of the programme [readers may gather that I’ve been slow to catch its denouement]

Fairness lies at the core of the concept of Squid Game, Netflix’s dark Korean drama. The players induced to enter the game feel they have no options in a highly unequal society, and have often been cheated out of what little money they did have, in legal or illegal ways, or been caught cheating others and now have debts to repay. They take every risk in order to gain the chance of walking away with enough money to change their lives. In contrast, the so-called ‘VIPs’ are bored by the extent of their wealth and need fresh expensive thrills to stay even partially engaged and interested. Both groups have been dehumanised by unfairness – it’s not by chance that the VIPs wear golden animal masks. The sole individual not dehumanised is the protagonist, Seong Gi-hun, who makes a particular point of insisting on his humanity as he departs from the scene.

However, as discussed in this blog’s original Squid Game ‘fairness’, the concept of fairness was considered only a little in the first series of Squid Game. Then, the organisers manipulated the idea of fairness as the basis to argue that it was right to punish those who had transgressed the rules of their own peculiar and dystopian world.

In the third series, the concept is deployed a little more overtly, particularly as part of a moving speech from the old lady, Jang Geum-ja (number 149) in the third episode, It’s not your fault. Jang deploys a cynicism about fairness that seems wholly appropriate to the cynicism of the programme overall:

“No matter how you look at it, life just is unfair. Bad people do bad things, but they blame others and go on to live in peace. Good people, on the other hand, beat themselves up about the smallest things.”

The organisers also again deploy the sense of fairness as a tool to justify their actions. In this case, they use it in praise of their own – again, peculiar and dystopian – version of democracy. Early in the first episode, the announcer asserts while the players view the rebellious dead from the end of series 2: “You are witnessing the fate of those who refused the democratic process of voting and instead attempted to stop the game using violent means. We will not tolerate any irrational behaviour which attempts to destroy the fair rules of this game, and such actions will be punished in accordance with our strict standards. We thank you again for your cooperation.”

The irony of complaining about the use of violent means to subvert democracy shouldn’t be lost on viewers: after all, there is an openness to, even a welcoming of, bullying and violence as part of and following the regular public ballots that deliver that ‘democracy’ (there’s a key moment when the organisers surprise us all by barring further violence between the players, at least outside the arena itself). The labelling of individuals by how they have voted fosters this scope for bullying and violence further: the series as a whole reminds us of the importance of the secret ballot to the effectiveness of democracy, to each voter being free to express their own opinion without unfair interference. In the absence of secret ballots, Squid Game’s democracy is a sham. As highlighted in Democracy – playing fair, fairness, equality and freedom are all necessary elements for true democracy to work. It’s clear to the viewer that the ‘democracy’ in Squid Game isn’t truly fair, in spite of the announcer’s words.

Fairness matters, deeply, in Squid Game; just not in the ways that the organisers claim.

I am happy to confirm as ever that the Sense of Fairness blog is a purely personal endeavour.

What’s fair pay for a bank?

It’s fair to say that generally banks aren’t well known for the fairness of their approach to pay. So it was interesting to note the emphasis that UBS, now by far Switzerland’s dominant bank following its 2023 rescue of the failing Credit Suisse, places on the issue of fair pay in its latest, recently issued annual report.

But the problem is that it seems to this reader at least that the bank may not necessarily be focusing on the right forms of fairness in pay.

For example, in the introductory section of the 43-page compensation report, the bank answers its own question, How does UBS support pay fairness?, as follows:

“We pay for performance, and we take pay equity seriously. Across all our locations, we apply the same fair pay standards, reinforced by annual reviews of our approach and policies in line with established equal pay methodologies. In 2025, our statistical pay gap analyses reaffirmed that pay differences between male and female employees in similar roles across our core financial hubs remained below 1%, a difference consistent with that for 2024. If we find any gaps not explained by business or by appropriate employee factors, such as role, responsibility, experience, performance or location, we look at the root causes and address them.”

UBS develops this analysis a little further in the Compensation philosophy and governance section, under the title Fair and equitable pay. Key elements of this discussion read:

“Pay equity and equal opportunity are fundamental to support our strategy. Being an employer of choice and inclusive of all experiences, perspectives and backgrounds is critical to our success. Factors such as gender, culture, race, ethnicity, sexual orientation and identity, disability, family, veteran status, generations and part-time status should not impact opportunities available to our employees.

“Fair and consistent pay practices are designed to ensure that employees are appropriately rewarded for their contribution.”

It again emphasises that gender pay gap analysis shows gaps below 1% in pay for male and female employees “in similar roles across our core financial hubs”, an interesting geographical narrowing and also clearly ignoring the general experience that the issues in gender pay arise most often because of differential opportunities for men and women, which tend to lead to a skewing of roles away from strict ‘similarity’. And note that it has a marked gender skew overall in staff, given that only 41% are female.

If we set aside these limitations, this discussion of fairness seems fine as far as it goes, but it barely begins to address the promise of the titles given to it. This is not UBS ‘supporting pay fairness’; this is simply the bank not overtly discriminating in its treatment of different staff – which in many countries and cases will in any case be illegal. An actual discussion of the challenge of supporting pay fairness might acknowledge realities outside the bank, rather than this internal focus. The financialisation of our economies has helped fuel the broad inequalities that our world faces, as financial institutions, competing with each other for ‘talent’, bid up the pay of individuals well beyond what is affordable or realistic (even imaginable) for most businesses. The closest UBS comes to a proper understanding of what a broader mindset about fair pay, looking beyond the financial sector, might actually imply is this comment, which at least acknowledges some external benchmark:

“We also aim to ensure that all employees are paid at least a living wage. We regularly assess employees’ salaries against local living wages, using benchmarks defined by the Fair Wage Network. Our analysis in 2025 showed that employees’ salaries were at or above the respective benchmarks.”

A bank that was genuinely seeking to ‘support pay fairness’ might apply similar pay expectations – of that minimum living wage level pay – at all clients and counterparties. A bank that worked with its corporate customers to understand what local living wages might be in their countries of operation and how they might be achieved across relevant workforces while still delivering profitability would be a bank genuinely supporting this ambition.

A bank that was genuinely seeking to ‘support pay fairness’ might consider whether financialisation is leading to pay distortions between the financial sector and almost every other part of the economy.

These UBS does not do. As ever, the finance industry has a tendency to look inwards at itself and not outward at its role and influence in the real world. An article I reference a couple of times in my forthcoming book invites the industry to think about fairness ‘outside its cocoon’. Only if financial services starts to do that will it genuinely be supporting fairness.

Duncan Mavin, in his (highly recommended) book telling the sad history of UBS’s takeover target, Meltdown: Scandal, Sleaze and the Collapse of Credit Suisse, identifies an unfair approach to pay as being part of the drivers of that bank’s failure: “The behaviour of the bank’s leaders hardly inspired other staff to be the best versions of themselves … Credit Suisse bankers got paid well, whatever happened. When the bank was making a profit, staff made a bundle, regardless of whether the results were driven by strong markets or great management. When the bank made a loss – because of misconduct, fines, bad behaviour or poor strategic decisions – the bonuses were good then too.”

Inward-looking financial institutions are more likely to fail than ones that look outward and measure fairness against external benchmarks. Instead, they need to look beyond their comfortable cocoon.

Annual Report 2025, UBS

Fairness Outside the Cocoon, Meir Statman, Financial Analysts Journal, Vol 60, No 6

Meltdown: Scandal, Sleaze and the Collapse of Credit Suisse, Duncan Mavin, Pan Macmillan 2024

Playing fair in detective fiction

A new novel called Fair Play was of course irresistible for this blog – especially when it turned out that the author, debut novelist Louise Hegarty, was speaking at my local bookshop.

Fair Play is an intriguing, even odd, blend. It starts as a modern story of a group of friends at a New Year’s party in an Airbnb country house. When one of the party dies mysteriously, it fractures into two: a locked-room detective story of the traditional sort, aping Agatha Christie, Arthur Conan Doyle, Dorothy L Sayers and so on, and the story of the grief of the dead man’s sister, Abigail, as she tries to make sense of what makes no sense.

The title of the book comes from what Hegarty has gathered as the ‘Fair Play Rules’ of detective fiction, three sets of guidance from its heyday of the late 1920s. Essentially, these are about being fair to the reader, so that we feel we have a fair chance of finding the truth, just as much as the detective, and at least we are smarter than the detective’s foil, the person Father Knox in one of the three sets of guidance, his 1929 Introduction to The Best Detective Stories of 1928-29, cruelly calls “The stupid friend of the detective, the Watson”.

It’s not hard to think of especially famous examples that breach some of the ‘Twenty Rules for Writing Detective Stories’ by SS Van Dine in The American Magazine in September 1928, for example numbers 12 and 13:

“There must be but one culprit, no matter how many murders are committed”; and
“Secret societies, camorras, mafias et al have no place in a detective story”

But then, some of the best writers have always played with the genre, on occasions bending the rules, only to the greater pleasure of the reader. The sleights of hand to pass quickly over clues after having brought them to our notice is one of the joys of these books (even if we only spot them after the fact!).

It’s not just my admiration for him as a writer that means my favourite of the three sets of guidance is that from TS Eliot, in a 1927 New Criterion piece in ‘Homage to Wilkie Collins’, widely seen to have invented the detective novel. The fifth of these is that “The detective should be highly intelligent but not superhuman. We should be able to follow his inferences and almost, but not quite, make them with him.” While being superhuman is not allowed, odd habits are: Hercule Poirot’s fastidiousness was clearly invented to hide (in plain sight) a decisive clue from the reader. It’s likely that same fastidiousness is what eventually came to annoy Christie so much about her character.

The key point of each of the guides is that the reader of a detective novel must emerge from the story not feeling cheated but feeling we could have got the answer for ourselves if only we had paid more attention, thought a little harder or taken a little more time to mull the clues available to us. We wouldn’t recommend to others a detective novel that was unfair by failing to live up to these expectations, and we probably wouldn’t read more from that writer. In reading, as in other things, humans favour fairness.

It needed no confirmation that Hegarty is a fan of detective stories, and of Christie in particular. Her handling of various of the standards of the genre, and particularly the playful repeated versions of different revelations of alternative murderers, show that very clearly. But in many ways the half of the book that is the detective novel is slight. What elevates Fair Play, and makes it linger in my mind, is the half that is the story of the sister’s grief. This, sparsely told and without easy answers, carries heft without being heavy.

The thing that seemed off limits at the talk was Hegarty’s own experience of grief. But it appears clear from the way she writes of it that she knows whereof she writes. We all have our experiences of grief, and this writing rings true, if anything helped by its sparseness. There are different short vignettes giving a vivid expression. Some of these are jarringly within a workplace setting, memorably a dull business meeting that comes to echo only with the words ‘my brother’s dead’. This feels very real, as does Abigail’s search for the answer, the simple revelation that will give her resolution. But life rarely offers the simple resolutions of a brilliant detective’s summing up.

Though there are no easy answers, I never felt cheated by what is an admirable first novel.

I am happy to confirm as ever that the Sense of Fairness blog is a purely personal endeavour. I am also happy to wish readers the compliments of the season.

Louise Hegarty, 2025. Fair Play, Picador

Father Knox, 1929. Introduction to The Best Detective Stories of 1928-29, Faber

SS Van Dine, 1928. Twenty Rules for Writing Detective Stories, The American Magazine

TS Eliot, 1927. Homage to Wilkie Collins: An omnibus review of nine mystery novels, New Criterion

Bubbles and economic fragilities

With all the talk about a bubble in investment in so-called ‘AI’*, I have taken a moment to reread the classic on the 1929 Wall Street bubble bursting, US economist John Kenneth Galbraith’s The Great Crash 1929, first published in the 1950s.

Speculative market bubbles do come and go. As Galbraith notes, the reason that 1929 is most remembered is not so much that the speculative bubble had grown so large before bursting (though it was unusually large) but that there were such broad real economy impacts from the bursting of the bubble – the lost years known as the Great Depression. A pair of Galbraith datapoints start to capture the scale of the Great Depression in the US and its searing impact on ordinary people: unemployment in 1933 was 13 million, one in four of the labour force; and even in 1938 still one in five were out of work.

So trying to understand why the Great Crash sparked the Great Depression is of real interest and seems like a timely thing to consider. Galbraith has no patience for the Wall Street apologists who argue there was no connection between Crash and Depression, but he does see that there were vulnerabilities in the economy that made it particularly susceptible to the crisis.

The first of these is of most interest to this blog, and Galbraith headlines it ‘The bad distribution of income’, noting that in 1929 the top 5% received around 35% of all personal income, and that interest, dividends and rental income (the almost exclusive preserve of the wealthy) represented fully 22% of total family income.

“This highly unequal income distribution meant that the economy was dependent on a high level of investment or a high level of luxury consumer spending or both. The rich cannot buy great quantities of bread. If they are to dispose of what they receive it must be on luxuries or by way of investment in new plants and new projects. Both investment and luxury spending are subject, inevitably, to more erratic influences and to wider fluctuations than the bread and rent outlays of the $25-week workman. This high-bracket spending and investment was especially susceptible, one may assume, to the crushing news from the stock market in October 1929.”

Readers will recognise some of our current distortions in these reports of the imbalances of pre-crash 1929 (also see Is enough enough? Addressing the problem of the super-rich, and The centre cannot hold). Significant inequalities – especially unfair ones – make economies less robust, more risky and more prone to crisis.

That is the first of Galbraith’s linkages between Great Crash and Great Depression. The others are as follows (deploying my brief characterisations of his comments):

Bad corporate structure. A business sector including many swindlers and fraudsters.
Bad lending. Profligate lending to unsound businesses and investments.
Imbalanced trade positions. Long-term trading imbalances, sometimes exacerbated through the application of tariffs, with the resulting deficits sometimes filled by corrupt, or at least grey, payments.
Poor economic insight. Shonky economic data riddled with holes.

So clearly, we’ve nothing to worry about now.

I will leave Galbraith with the last words of this blogpost, without comment from me. In the last pages of the book, he writes: “during the next boom some newly rediscovered virtuosity of the free enterprise system will be cited. It will be pointed out that people are justified in paying the present prices – indeed, almost any price – to have an equity position in the system. Among the first to accept these rationalizations will be some of those responsible for invoking the controls. The newspapers, some of them, will agree and speak harshly of those who think action might be in order. They will be called men of little faith.”

* Seasoned readers may remember that I am an ‘AI’ sceptic – see A just AI Transition, for example

I am happy to confirm as ever that the Sense of Fairness blog is a purely personal endeavour – and also that I do not give, and am not authorised to give, personal financial advice. This blogpost should not be construed as such advice.

John Kenneth Galbraith, The Great Crash 1929. Hamish Hamilton 1955

The power of powerlessness

From this week’s Economist: how our sense of fairness means that we favour the underdog, and so makes forms of protest that seem powerless, powerful.

[if anyone is missing my blog, don’t worry! My fairness writing energies are largely being taken up by a book commitment; I’ll return re-energised to the blog next year, and you’ll also be able to read a book version of some of my Sense of Fairness thoughts]

Workers value dignity

It shouldn’t really be a surprise, but dignity at work – a combination of things such as the sense of autonomy and relationships with colleagues and bosses, and being treated fairly – matters to people. It’s as true at the bottom of the income scale, where observers might assume concerns about pay outweigh all other considerations, as it is higher up. Dignity matters to people, as I’ve been exploring in recent blogs.

For the book I am writing (on fairness in business and investment) I am currently investigating the literature on monopsony and oligopsony in labour markets. Monopsony is the distorted market situation arising from there being a single buyer of a good or service (a monopoly is where there’s a single seller); oligopsony is where there is a narrow enough group of buyers that they distort the marketplace. Economists are increasingly observing evidence that the labour market suffers inefficiencies that are consistent with oligopsony – employers having excess power in setting pay. Most workers would probably agree that their experiences too are consistent with this.

One part of this literature particularly stood out because it made a link to the issue of dignity, which increasingly seems a key element of people’s innate sense of fairness, and of their inclusion in society and the economy. In particular, a 2022 paper from the US National Bureau of Economic Research, called Power and Dignity in the Low-Wage Labor Market: Theory and Evidence from Wal-Mart Workers, uses evidence from real interactions with US employees of the globe’s largest private sector employer to understand their views and test hypotheses against reality.

The results are striking.

The study included four sentences exploring the degree to which workers had a sense of dignity in their jobs (these sentences were developed based on prior interviews with Wal-Mart workers that sought to understand their experience in the workplace, as well as earlier academic work). The overarching question was Indicate to what extent the sentence describes the workplace of your job at Walmart, and each time respondents were offered four responses (Almost Always; Often; Sometimes; Never). The four sentences were:

You [have/had] the opportunity to express yourself while at work.
You [can/could] rely on your co-workers to help you with work.
Your supervisor [treats/treated] you with respect.
Your supervisor [treats/treated] everyone fairly.

And of these four measures of dignity, it appears to be fairness that matters most. Indeed, a lack of fair treatment by one’s boss is in essence the greatest determinant of likelihood of quitting a job in the study, with the obvious exception of pay (and of the availability of hours of work a week, which is a clear part of the pay equation for those paid on an hourly basis):

Consistently, the study confirms that fairness and dignity are powerful drivers of work satisfaction, and thus in willingness to stay with an employer.

As the study states:

“A natural question is whether firms can adjust the level of dignity at work. While immediate supervisors likely have the most discretion over workplace dignity, supervisors can be incentivized by higher-level managers to treat subordinates fairly and with respect, and workplace rules can be designed to allow opportunities for self-expression and co-worker support. While it may take time to alter workplace experiences, and agency costs might be considerable, the significant cross-store variation we document below suggests that managers have some control over the level of workplace dignity.”

Our bosses, and how they treat us, matter.

As well as enhancing people management, the authors raise the interesting challenge of whether improving the competitive context of the labour market is necessary to increase dignity in the workplace, the experience of fairness for workers:

“any effort to increase workplace amenities (including subjective experiences at low-wage jobs) may require policies that reduce monopsony power in the low-wage labor market. The high levels of labor market competition in the immediate post-COVID labor market may have given workers the opportunity to quit jobs that didn’t provide dignity. Whether this results in firms upgrading the subjective experience of work remains to be seen.”

I’m not sure that we’ve yet seen significant enhancements to workplace dignity and fairness, but perhaps we should continue to live in hope.

I am happy to confirm as ever that the Sense of Fairness blog is a purely personal endeavour

Power and Dignity in the Low-Wage Labor Market: Theory and Evidence from Wal-Mart Workers, Arindrajit Dube, Suresh Naidu, Adam Reich, NBER Working Paper No. 30441, September 2022

Belonging, not belongings

The modern mind profoundly misunderstands our ancient ancestors. Take the mystery of ancient stone circles. One of the oldest is the Ring of Brodgar on Orkney, probably a creation of the culture that later built Stonehenge. Most guidebooks assert that the 21 currently standing stones (8 of them re-erected between 1906 and 1908) are the remnants of a circle of 58 or 60 stones.

The trouble is, this common assertion just isn’t true. There is no sign that there were ever more than 40 standing stones in the Ring – the latest archaeology, using a range of techniques, cannot find evidence for more than 16 sockets for stones other than those still standing. The modern mind has extrapolated from the stones that remain and calculated how many there would be if there had once been a circle of that circumference made up of stones with similar gaps between them (and ignoring the fact that the gaps vary). But that wasn’t the point for the ancient mind. Our ancestors thought differently from us, sometimes in ways that are even odder to our modern eyes: some stones weren’t ever deeply planted in the ground and so may have fallen very quickly.

The point is that the Ring of Brodgar wasn’t a thing, it was a doing: for our ancestors it was the collective action that mattered, its creation, not so much the resulting thing. Brodgar was a community event that brought people together in a shared effort. The stones aren’t matching: they come from different spots around the islands (at least 7), presumably transported by different families the miles from their best quarry, and raised as part of a communal effort. The ditch around the stones was certainly dug in small patches that were then joined together, perhaps further evidence of the bonding process of each family doing its own portion of the work before they were brought together as a unified whole – overall a work requiring an estimated 80,000 person hours.

The Ring of Brodgar, both stones and gaps between them

It’s like Peru’s ancient Nazca lines, about which modern people, flying over them in aeroplanes, puzzle as to the creation of images that can only be seen from above in this way. Most archaeologists now agree that the Nazca images were ceremonial paths, walked by people on special days or for particular reasons. They were not meant to be seen but to be experienced, perhaps to be made anew. The Nazca lines too were a doing, not a thing.

I keep coming back to these thoughts as I read a thoughtful and thought-provoking recent book, Don’t Talk About Politics. Author Sarah Stein Lubrano notes that we fail to persuade each other on politics and are wasting time and energy in talking at – and past – people with whom we disagree. Her core point is that politics is a doing, it is something that we need to participate in rather than see as separate to us, as something about which we can debate while remaining dispassionate observers: “To think effectively about politics…is not only to think about action; it is also, by necessity, to take action. Good political reasoning cannot happen simply by reading and debating.”

She goes on:

“When it comes to politics, we’ve probably placed too much societal emphasis on words and too little on action. Many people would say that democracy is about whether people can make up their own minds and then vote. But this is, in fact, a fairly passive idea of political involvement…our collective common sense of what politics is has, I’d suggest, become hollowed out. It involves too little of what truly changes people, their priorities and their actions.”

Stein Lubrano’s focus is politics, but society too is a doing, and so is the economy. We need to be active participants in both, and dangers lie where many members of a population are not. And at present, too many people do not feel like active participants in either the society or the economy.

This thought process echoed for me across a recent discussion co-hosted by my friends at the Fairness Foundation. They’ll write more fully and directly on what was a discussion under the Chatham House rule, so I will stick with my reflections. The core of the conversation centred on the issues I touched on in An Inequality in Dignity, or the Dignity Deficit; one participant talked about there being a ‘crisis of belonging’ that we all need to seek to address actively.

We discussed the way in which many people currently feel profoundly uninvolved in the economy. It isn’t just that they own little and lack a stake in the economy; more, they feel that the economy is something that occurs elsewhere in which they are not engaged. They lack both belongings and belonging. The unfairness embedded in the current economy, where so many are have-nots, is profoundly alienating and destructive of human dignity. Similarly, a dignity deficit arises because people do not feel they have a place in society, that their voices are not heard and will not be heard, that they have no agency. They withdraw into fury and anger because they feel that they no longer have a place in which they are welcomed as belonging.

This is true in the economy and in society more broadly, and the tensions and disconnects it fuels are plain to be seen. Stein Lubrano devotes a whole chapter to the psychological side of these sentiments. She refers to it as social atrophy, a sense that everyone is becoming isolated, detaching themselves from society and community. She notes the disastrous consequences that arise. As she points out, this is a much more fundamental risk than the way it is usually discussed, as an individual crisis of loneliness; it is a broader communal fracturing. “We may…distrust society more not only because it no longer provides everyone with what they need, but also because, as we become more isolated, the suspicious and paranoia-inducing structures in our brains become more and more activated,” she argues.

If we are to address the dignity deficit, the crisis of belonging, we need to involve people, we need for them to feel that society and the economy (and yes, politics too) is something in which they have a role to play, that these are not just issues that happen elsewhere between other people, but rather that they can act and be engaged. That will require more physical venues for people to meet face to face, for society physically to thrive. It will require a democratisation of finance, so that ordinary people understand that they do literally have a stake in the economy and are economic actors in their own right – an understanding that their pension savings make them part-owners of assets and of companies around the world. This needs to be understood as giving them agency, the scope to be involved and to act, not simply to be onlookers at activity involving others.

It turns out that our ancestors were right: society is a doing, it isn’t about things. We need to get out of the trap of the modern mind, that emphasises the thing rather than the doing. We need less emphasis on belongings and more on belonging.

I am happy to confirm as ever that the Sense of Fairness blog is a purely personal endeavour

Landscapes Revealed: Geophysical survey in the heart of the Neolithic Orkney World Heritage Area 2002-2011, Amanda Brend, Nick Card, Jane Downes, Mark Edmonds, James Moore, Oxbow, 2020

Building the Great Stone Circles of the North, Colin Richards (ed), Windgather, 2013

Don’t Talk About Politics: How to Change 21st Century Minds, Sarah Stein Lubrano, Bloomsbury Continuum, 2025

What’s a fair use?

The media has reported that Meta (the Facebook, Instagram and WhatsApp company) has won a legal case on the use of copyrighted materials in training its AI models, that the use of copyright materials was a ‘fair use’. As often with the law, it’s a bit more complicated than that.

The case in question was Kadrey v Meta, and summary judgement was released last week (the judge, Vince Chhabria, deciding on the basis of arguments that the case did not need to go to jury trial because the plantiffs had not made a convincing case, enabling Meta to succeed in a call to dismiss it). The legal question at issue was whether the accepted abuse of copyrighted works in training AI amounts to a ‘fair use’. As well as considering fairness, the case opens a wider window on AI.

Before delving, I will note that I’ll continue to use the term AI, because it’s used in the case and the term is in general use for these emerging new technologies. But as both recent books The AI Con and AI Snake Oil (the two latest additions to my bookshelf) start off by making clear, there is no such single thing as AI. It is a catch-all term for a range of technologies – some of only very dubious effectiveness – and is really just a brand that is being deployed to raise (enormous amounts of) funding (two headlines from the Financial Times over this weekend cast light on the scale of this financing: Meta seeks $29 billion from private credit giants to fund AI data centres, and Nvidia insiders cash out $1 trillion worth of shares). The best known, and most used, of these new AI technologies are called large language models (LLMs), accurately described as stochastic parrots: models that simply put one word after another according to statistical models developed through their training.

Many legal systems favour the term fairness, and ‘fair use’ is a well-established concept in US law. The country’s Copyright Act (in 17 USC §107) clearly restricts fair use to usage “for purposes such as criticism, comment, news reporting, teaching (including multiple copies for classroom use), scholarship, or research”. It sets out four factors that should be considered in determining whether a given use is in fact fair:

1. the purpose and character of the use, including whether such use is of a commercial nature or is for non-profit educational purposes;
2. the nature of the copyrighted work;
3. the amount and substantiality of the portion used in relation to the copyrighted work as a whole; and
4. the effect of the use upon the potential market for or value of the copyrighted work.

Deciding what uses are fair is both a matter of law and of the specific facts, meaning that there are multiple cases that have considered these factors. The list of four factors is not exhaustive, but are assistants in reaching the overall conclusion. The fourth factor, whether the use risks substituting for the copyright materials in the marketplace, is generally seen to be the most important. Courts need to apply judgment and consideration is deciding on fair use; as ever, assessing fairness requires thought and judgment.

As judge Chhabria explains in his summary judgement:

“What copyright law cares about, above all else, is preserving the incentive for human beings to create artistic and scientific works. Therefore, it is generally illegal to copy protected works without permission. And the doctrine of “fair use,” which provides a defense to certain claims of copyright infringement, typically doesn’t apply to copying that will significantly diminish the ability of copyright holders to make money from their works.”

He is as rude as a judge ever gets about a fellow judge who reached a recent decision on a fair use case in relation to Anthropic, another AI firm (Order on Fair Use at 28, Bartz v Anthropic PBC, No. 24-cv-5417 (N.D. Cal. June 23, 2025), Dkt. No. 231). That judge was convinced by the argument that training AI was no different from – and had no more impact on the market for copyright products – than training schoolchildren to write. Chhabria says: “when it comes to market effects, using books to teach children to write is not remotely like using books to create a product that a single individual could employ to generate countless competing works with a miniscule fraction of the time and creativity it would otherwise take. This inapt analogy is not a basis for blowing off the most important factor in the fair use analysis.”

And surprisingly given his overall ruling, Chhabria is very clear that AI companies are breaching copyright law and are damaging the commercial market for copyrighted works. He seems very sure that AI companies fail at the fourth factor in assessing fair use: “by training generative AI models with copyrighted works, companies are creating something that often will dramatically undermine the market for those works, and thus dramatically undermine the incentive for human beings to create things the old-fashioned way”.

Chhabria also notes a simple flaw in one of the AI companies’ arguments: that applying copyright law will stifle the development of this technology. He notes that any finding that this use of copyrighted materials isn’t fair use does not bar that use, it just requires that AI companies need to reach a commercial agreement with copyright holders to compensate them for the – unfair – use of their materials. As he points out, these businesses project that they will make billions, indeed trillions, of dollars from AI services, so should be able readily to afford such licensing. Indeed, the court saw evidence that Meta initially sought to licence book materials for training purposes, and considered spending up to $100 million on doing so. This never happened because book publishers do not hold rights to this use of book materials – like other novel uses, the rights rest with the authors – so there is no central point or points for such a negotiation. The fact that AI companies are seeking direct commercial benefit from their use of copyright materials makes their burden in demonstrating fair use much harder.

Despite Chhabria’s conclusions that seem to strongly favour the copyright-holders who brought the case, he nonetheless found against them. The copyright holders are 13 authors who argued that their works had been used in training Facebook’s Llama LLM models. In essence they failed in their claim because their lawyers focused their efforts and arguments in the wrong place. They made their arguments predominantly under the first three of the four factors in §107 of the Copyright Act, and failed in those. While the fourth factor – the effect of the use on the potential market for the copyrighted work – is generally seen as the most important, that is not an argument they made strongly. They simply did not argue (or were at best “half-hearted” in those arguments) that their works had been used as the basis for a tool which might flood the market with similar works, undermining the value of their copyright, nor did they provide evidence to support such an argument. This was “the potentially winning argument” according to Chhabria; the (weaker) points actually deployed in argument before the court did not succeed.

Chhabria was clear:

“this ruling does not stand for the proposition that Meta’s use of copyrighted materials to train its language models is lawful. It stands only for the proposition that these plaintiffs made the wrong arguments and failed to develop a record in support of the right one.”

It does seem ludicrous that the most valuable companies in the world should argue that it is fair for them to take stolen copies of books subject to copyright protection (the training materials were taken from so-called ‘shadow libraries’, of illegally scanned books) and make what they predict will be huge commercial profits as a direct result, while providing the copyright holders with no compensation. The fact that Meta explored licensing but found it too difficult and delaying helps support the case that this would be the right thing to do.

The Kadrey case reports one other specific element of the training of Llama models – that they are taught not to produce more than 50 words together that are repeated from any one source (even if provided with highly directive prompts to do so). The fact that this is a deliberate part of the training shows just how prone these technologies are just to leaning on what they have read. In a recent Financial Times interview, Professor Emily Bender, coiner of the term stochastic parrots and co-author of both the academic article that brought the term to prominence and of The AI Con, is quoted as calling LLMs “plagiarism machines”.

I have to admit that, as may be apparent from my recent reading habits, that I am an AI sceptic. I suspect that we will look back on this period with puzzlement, and wonder why we threw colossal amounts of computing power – and colossal levels of energy in our carbon-constrained world – at jobs that human brains are better at. AI is neither artificial nor intelligent: it isn’t artificial because it depends on human creativity in the training, and it also depends on significant, horrible, labour (typically cheap precarious labour in emerging economies) in cleansing the models of the filth that it produces because it has been trained on, among other things, the global sewer that is the Internet. It isn’t intelligent, it’s just reproducing others’ language patterns based on statistics, “haphazardly stitching together sequences of linguistic forms it has observed in its vast training data…without any reference to meaning” as the stochastic parrots paper put it. As Bender told the FT, we are “imagining a mind behind the text…the understanding is all on our end”. There will no doubt be jobs that AI technologies are useful for, but like any human tool it is tailored to its task, and not a general purpose vehicle for all activity. Currently we have a hammer and are making the mistake of seeing everything as a nail.

As a result, I suspect that much of the billions being deployed in AI currently will turn out to have been wasted. I should admit also that my view may be coloured by the fact that I entered the investment world exactly at the time of the dotcom bubble. While I avoided losing money in the dotcom bust, I also missed out on investment gains as that bubble inflated.

But this is a blog on fairness, not AI cynicism. The Kadrey decision did not conclude that Meta’s actions were fair, only that the copyright-holders had failed to deploy the arguments that might have shown how unfair the use of their materials was. This will clearly not be the last such case, and while the AI businesses will continue to deploy some of their investors’ millions into their defence, judge Chhabria’s legal conclusions suggest they will have a challenging time winning cases argued on the right basis.

Rather than finding that Meta’s use was fair, the Kadrey decision is highly suggestive that AI is not fair in its use and abuse of copyright materials. That feels right: fairness should always tend to rebalance power away from those with billions towards those of whom they take uncompensated advantage.

I am happy to confirm as ever that the Sense of Fairness blog is a purely personal endeavour

Kadrey v Meta, Case No. 23-cv-03417-VC, Summary Judgement 25 June 2025 (Docket Nos 482, 501)

The AI Con: How to Fight Big Tech’s Hype and Create the Future We Want, Emily Bender, Alex Hanna, Bodley Head, 2025

AI Snake Oil: What Artificial Intelligence Can Do, What it Can’t, and How to Tell the Difference, Arvind Narayanan, Sayash Kapoor, Princeton University Press, 2024

Meta seeks $29 billion from private credit giants to fund AI data centres, Eric Platt, Oliver Barnes, Hannah Murphy, Financial Times, 27 June 2025

Nvidia insiders cash out $1 trillion worth of shares, Michael Acton, Patrick Templeton-West, Financial Times, 29 June 2025

The Copyright Act, 17 USC

AI sceptic Emily Bender: ‘The emperor has no clothes’, George Hammond, Financial Times, 20 June 2025

On the Dangers of Stochastic Parrots: Can Language Models Be Too Big?, Emily Bender, Timnit Gebru, Angelina McMillan-Major, Shmargaret Shmitchell, Proceedings of FAccT 2021