Learning from the stochastic parrots: finding fairness in AI

“we advocate for research that centers the people who stand to be adversely affected by the resulting technology, with a broad view on the possible ways that technology can affect people”

It doesn’t sound like a particularly radical statement, certainly not radical enough to warrant being sacked. It says no more than the standard ethical question, just because we can, should we? Yet this seems a radical question to some, and indeed two of the authors of the paper from which this quote comes have been sacked by their employer from top roles in ethics, seemingly for reasons closely connected to the paper (though the exact circumstances and reasoning remain a matter of dispute). 

Those former employees are Timnit Gebru and Margaret Mitchell (not really masked by being named as Shmargaret Shmitchell in the list of authors of the paper), and their former employer was Google, known to investors as Alphabet. There are also three further anonymous authors of the paper, prevented by their employer (it is specifically stated as a single employer) from being named; most assume that that employer is also Alphabet/Google.

The article at the heart of the dispute is On the Dangers of Stochastic Parrots: Can Language Models Be Too Big? [it also includes a parrot emoji as part of its title]. Stochastic Parrots is a short article, with a substantial catalogue of references, which contains few surprises. The Data Ethics Club in which I participate could see nothing at all that was truly controversial, but then we are perhaps a self-selecting crowd. Stochastic Parrots was prepared for the FAccT conference – the Association for Computing Machinery’s Fairness Accountability and Transparency conference due to start this coming week (it is to be discussed on March 10th, the last day of the conference) – and has been published by Emily Bender, a University of Washington academic who is another of the co-authors. The biggest surprise for most is that it is controversial enough to spark these departures. 

The article points out that AI can use significant amounts of energy, and in a world of CO2 constraints, technology firms should not regard the use of endless amounts of data to train new algorithms as entirely cost-free. Again, it notes that the current trend for supposedly ‘natural’ language programs being trained on the entire Internet encodes the existing biases and unfairnesses built into current web activity, and further extends the domination of heavy users of tech while excluding those who are not part of the Internet community, those we might call ‘unwebbed’. It is like all the world were expected to adopt a West Coast US accent and verbal mannerisms – as if!

Both these concerns are clearly true and while they may be a little embarrassing for Big Tech they cannot be truly controversial.

The article goes on to criticise the large language programs. These, it argues, do not generate meaningful outputs but merely sewn-together repetitions of language they adopt from elsewhere, joined not by meaning but through statistical habit. They are, in short, the stochastic parrots of the title. Now this is a well-articulated criticism of a large area of activity for Google and its Big Tech peers, and the description (and emoji) makes it memorable, but it is hard to see it as a sackable offence for those charged with considering ethics to note that a work in progress has distinct weaknesses and doesn’t yet live up to the promises made for it.

So I suspect that it is the last aspect of Stochastic Parrots that is the genuinely controversial part, the raising of the fundamental ethical question of just because we can, should we? The fact that it was perceived as controversial reveals much about Google and Big Tech.

As well as the quote that heads this blog, Stochastic Parrots goes a little further: “we call on the field to recognize that applications that aim to believably mimic humans bring risk of extreme harms. Work on synthetic human behavior is a bright line in ethical AI development, where downstream effects need to be understood and modeled in order to block foreseeable harm to society and different social groups.” 

So, it seems that Big Tech – or at least Google/Alphabet – is unwilling to countenance the very idea that some of its activities might have consequences that amount to extreme harms to people. And they seem to be unwilling to accept the idea that they should adopt a cautious approach of assessing what those harms might be before proceeding with their activities. The irony that these comments are very close to an articulation of the ‘don’t be evil’ mantra that framed Google at its creation is clear. It is also clear that ‘don’t be evil’ seems not to be applied by Alphabet now.

Such a denial of the possibility of risk and of the need to be cautious in the face of it seems a very dangerous position for such a powerful industry to take. And it flies in the face of the readily apparent harms from AI: perhaps it is no wonder that we are seeing such harms given the state of denial in the industry.

Ethics in AI remains a new field, worryingly. The overwhelming thinking does seem to be not to ask whether we should do it just because we can, but to do it simply because we can. A recent article in the New Yorker, Who should stop unethical AI?, sets out the trauma of beginning to consider ethics in AI. One quote, from an individual who was then an academic researcher but now works at Google, seems particularly telling: “Obviously, researchers are incentivized to pretend that there are no human subjects involved, because, otherwise, things like informed consent become issues, and that’s the last thing you want to do when you’re processing data with millions of data points.”

And there is a risk that ethics in AI may never get beyond its nascent phase. As Bender points out in an interview with MIT Technology Review, the experience over this paper may hinder further discussions of ethics in AI – it will have a “chilling effect” she is reported as saying. Having AI ethics specialists working for the Big Tech firms “has been beneficial in many ways. But we end up with an ecosystem that maybe has incentives that are not the very best ones for the progress of science for the world.”

Yet we all know that we need ethics in AI. We know that Big Tech has facilitated, indeed fuelled, a polarisation in politics, and society. The algorithm is the echo chamber. Big Tech has eroded privacy and increasingly exploits its knowledge of us to direct our consumer choices. As Stochastic Parrots worries, Big Tech risks reinforcing, not unwinding, existing unfairness in our society. 

So we need ethics in AI. And part of that has to be more transparency. There is a huge irony in the fact that Big Tech is among the most secretive businesses in the world. It is ironic because largely it insists on no one else having secrecy, and thrives on undermining other intellectual property while keeping its own entirely hidden. The industry has at best a lax attitude to other’s copyright, whether those others are individuals or traditional media outlets (as has been seen most recently in the Facebook and Google spat with Australia). Big Tech has an avaricious attitude to every individual in the world’s personal data, repurposing it as the basis for a uniquely targeted advertising model.

Big Tech has even played a major role, according to Rana Foroohar in her searing Don’t be Evil: The Case Against Big Tech, in undermining the protection offered by the patent regime in the US. Since the passage of the America Invents Act in 2011, she reports, the US regime has swung from one of the easiest to gain a patent and to protect it to one where such protection is much harder. 

Yet Big Tech also has an obsession with maintaining the privacy of its own organisations and of hiding the workings of its core business model, the algorithm, from the outside world.

Academics are used to testing their work in the outside world: in a sense, it does not exist unless it has been peer reviewed and published openly. Similarly, patents allow their owners unique benefit from their creativity for a period of years, on the basis of being made publicly available so that the knowledge can be freely available in due course and built upon by others. Open source software, which is publicly available and so can be tested and improved by developers around the world, is generally regarded as much more robust than secretive proprietary software: for example, it is not by chance that NASA chose Linux and other open source software to be the brains of Ingenuity, the drone currently flying itself in Mars’s scanty atmosphere.

Patent law developed because many inventions are easily replicable. New products can be reverse engineered and therefore copied, and society agreed that in order to foster creativity inventors need to be able to prevent such copying for enough time to generate a return from their invention. But the price of such exclusivity is publicity: the patent is made public and thus is freely available after the exclusive period ends. It can also be subject to testing and enquiry, and can be built on for future inventive leaps. A patent attorney friend notes that one of his first questions to a client or potential client is always do they actually want a patent at all. If reverse engineering isn’t possible so a secret cannot be readily discovered by the outside world, would it not be better to keep it as a secret and never release it to the public through the patent process? Coca Cola has proved over more than a century the value of a commercial secret that is kept as a secret. It is this model that IT giants – ironically given their obsession with freely revealing all the world’s information – are using.

I have previously lauded CEO of the British Academy Hetan Shah’s idea that there should be a data commons: that our data should not be owned by any business but rather available for common usage, perhaps after a period of commercial benefit (like the time-span of a patent) to enable the garnering of the data to be funded. But perhaps we should be brave enough to go further than this: all algorithms that affect more than a minimal number of people could be made subject to independent oversight by some properly funded regulator and accessible for academic research. This would enable others to test and challenge its quality, and to identify glitches and biases – in particular, to spot potential negative impacts of algorithms on society and individuals. It could prevent harms being hidden behind the veil of commercial interest.

Shah proposes that proprietary rights to data that is gathered should expire after 5 years, after which the data should be released to a charitable corporation and made available to all for research and study, “subject to scrutiny that ensure the data are used for the common good”. This argument is particularly strong, as Shah points out, for health data and other broad data sources that governments should never lazily allow to be the everlasting property of a single corporation.

Sunlight is the best disinfectant, we know. But little sunlight is currently allowed into the world of AI. These are black boxes from which light and attention are excluded.

Transparency is a protective model that has been attempted, to some extent at least, and in Europe at least. For the heart of some of the protections offered to consumers by the EU’s General Data Protection Regulation (universally known as GDPR) are the underlying principles of lawfulness, fairness and transparency. Under transparency, data processors are expected to be open about themselves and how they will use individuals’ data – this is often expressed as the individuals’ right to be informed.

Of particular relevance is Article 22 of GDPR, which protects individuals from abuse through solely automated decision-making. Such processing is only permitted in limited circumstances, including after having obtained explicit consent. Individuals are also given rights to challenge and insist on human intervention in the decision-making, and the processor is obliged to carry out regular checks to make sure that the automated decision-making is working as intended. All very sensible and fair.

This Article formed a key element of a recent British Institute of International and Comparative Law and Kings College London (KCL) Dickson Poon School of Law event on Contesting AI Explanations. Essentially, the key message of this fascinating session was that GDPR only goes so far, and that this is not far enough. The fact that Article 22 allows challenge only of solely automated decisions limits matters as few decisions as yet are solely automated. More significantly, though, the challenge can only be made where there is enough visibility of the AI activity to allow insight and challenge. It also happens that GDPR does not require disclosure of the detail of code, but of ‘meaningful’ information, a concept which is largely untested.

The comments of Robin Allen QC, of Cloisters chambers and co-founder of the AI Law Hub, were typical. He noted that transparency is helpful but often the use of AI is not visible so that it cannot be challenged. Take his example of a CV uploaded to the Internet; this may lead to interviews and job offers, but the individual will be unaware of all the occasions when an algorithm has screened them out for what might be arbitrary or wholly unacceptable reasons. Unless it is visible that this has happened, all current protections count for nothing.

The conclusion of most participants was therefore that GDPR does not in practice provide the protections that we might hope for. Rather, standard public law challenges and remedies are more often sought – such as judicial review. Some argue that this is beginning to amount to a presumption of disclosure for government and public bodies. However of course, such challenges are only possible where the AI is employed by public bodies, and do not exist in relation to the private sector. 

We are left not just with holes in the application of the regulation, but with certain key conversations not even begun. The comments of Perry Keller, a reader in media and information law at KCL, were particularly striking. “We have not had enough of a debate about these harms,” he said. He noted that as well as the individual harm, there is a collective harm, a societal one, though GDPR considers only the individual aspects. “The excessive focus on the individual is not allowing us to focus on the societal one.” As Stochastic Parrots invites us to, we need to consider foreseeable harms, and explore whether their existence makes the new technology worthwhile.

We seem to have come a long way from don’t be evil. I can find no better ending than the first paragraph of Don’t be Evil: The Case Against Big Tech:

“’Don’t be evil’ is the famous first line of Google’s original Code of Conduct, what seems today like a quaint relic of the company’s early days, when the crayon colors of the Google logo still conveyed the cheerful, idealistic spirit of the enterprise. How long ago that feels. Of course, it would be unfair to accuse Google of being actively evil. But evil is as evil does, and some of the things that Google and other Big Tech firms have done in recent years have not been very nice.”

See also: The failures of algorithmic fairness

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

Data Ethics Club

Fairness Accountability and Transparency Conference

Who should stop unethical AI?, Matthew Hutson, New Yorker, February 15 2021

We read the paper that forced Timnit Gebru out of Google. Here’s what it says, Karen Hao, MIT Technology Review, December 4 2020

Don’t be Evil: The Case Against Big Tech, Rana Foroohar, Random House, 2019

Use our personal data for the common good, Hetan Shah, Nature 556 7, 2018

Contesting AI explanations in the UK, BIICL and Kings College London, 24 February 2021

AI Law Hub

Government Models and the Presumption of Disclosure, Joe Tomlinson, Jack Maxwell, Public Law Project, July 2020

Is fairness an art?

One of the main attractions of the concept of fairness is that it requires judgement, it is qualitative not quantitative. It is not reduceable to a number, and hence differs from inequality, the hard-edged term that tends to be favoured by most economists – because they mostly like to have a calculation and favour quantification. Measuring it doesn’t make inequality the better measure: the simple fact is, humankind can stand a good deal of inequality if it is fair; if it isn’t fair, equality is deeply unsettling. 

Fairness is a clear sense that is innate in humans and crucial to our well-being. But does that make it an art?

It’s a question forced upon us by the title of an excellent new book The Art of Fairness: The Power of Decency in a World Turned Mean, by David Bodanis (author of bestseller e=mc2), published late last year. I am happy to recommend the book.

The Art of Fairness is alive with stories, finding fair and decent behaviour around the world and across history, as well as finding its opposites. I am now sorely tempted to do further research on the Bounty’s Captain Bligh, who appears to have swung between periods as a generous and inspirational leader and as a vicious monster. And I have to admit favouring any book that has a chapter on the experience of the second world war in India and Burma, and the leadership of the great and largely forgotten General Slim, under whom one of my grandfathers served (in a very junior capacity). It remains an oddity that the British obsession with that war almost entirely neglects the war in Asia. 

From a business reader’s perspective, the two key chapters are one that contrasts Satya Nadella’s oversight of Microsoft with that of his predecessor Steve Ballmer, and another that compares the leadership that delivered the remarkably efficient construction of the Empire State Building with the venality that took apart an airline business, Eastern Airlines, apparently for personal greed. For me it felt that the Microsoft story held few surprises, as the downs and ups of that company in the years since Bill Gates’s departure have been played out in public, and the weaknesses of Ballmer seemed all too apparent to the outside world in his mostly mocked performances seeking to inspire staff, where he appeared unable to persuade even himself. 

The Empire State Building and Eastern Airlines stories, while having some familiarities, are less known and so it felt they had more to teach. The Empire State story focuses on lead contractor Paul Starrett, who had what we might now call a strong stakeholder approach to business. In particular, he saw significant value in paying workers above the standard rate and offering them better treatment, not least making decent food available to them with subsidised restaurants created on a number of floors as the building rose. He was no soft touch: the iron fist within this velvet glove was a rigorous audit process, for people, tools and materials – so that no one was tempted to skim additional wages for non-existent workers or scam materials away from the building site. Both were problems typical of the construction business then as now. Starrett was rewarded by the construction being swifter and more efficient than rival sites even given the scale of the ambition that the building represented. It remained the tallest building in the world for decades.

The Eastern Airlines story is a contrast to this. Frank Lorenzo became CEO of the successful business in 1986. He took over a thriving and successful operation with good stakeholder relationships. And he proceeded to damage each of those relationships in order to make more money. He cut salaries among the workforce. He sold off assets, including the innovative reservations system that the business had carefully built. It was worse than just selling assets: often these were related party transactions that benefited him personally. He tried to divide and rule the workforce, trigger a strike among ground crew and mechanics and enter bankruptcy to offload liabilities. His plans failed and he was ousted, but not before having so damaged the business that it was split up among a number of acquirers.

One of the familiarities of this Eastern Airlines story is that it feels like a good deal of current business behaviour: too often it seems companies now are being overloaded with debt and, if they do not succeed quickly, deliberately put into bankruptcy to enable a phoenix business to emerge unburdened by former liabilities. Bodanis, in his generous discussion of further reading materials and explorations that provides a final two dozen pages of the book, makes an explicit link between Lorenzo’s behaviour and that of the worst activities of private equity. This unfortunately feels a fair comparison.

Each of these chapters in the first half of the book is used as evidence for what Bodanis calls koans, Zen Buddhism’s brief, sometimes apparently contradictory lessons. Examples are ‘Listen, without ego’ and ‘Give, but audit’. Collectively, these koans build through the book to form Bodanis’s understanding of what the art of fairness amounts to.

The second half of the book is given over to a contrast between the always flawed but ever-determined US President Franklin Delano Roosevelt, and the flawed and irresolute Nazi propagandist Joseph Goebbels. Though both feel like stories we have heard before, Bodanis manages to vest each man’s history with fresh insight – framed by the koans evidenced by earlier chapters – helping us to learn anew.

While Bodanis does not explicitly draw links between his discussion of Goebbels’ big lies and current political tendencies, those messages are clear between the lines (and between the lies). Reading the book over the weeks either side of this January’s insurrection in Washington DC, as I did, it felt disturbingly prescient. Further, the discussion of FDR feels particularly important given that President Biden, indeed all of us, seem to be some form of new New Dealers now, in our determination to Build Back Better – or better, to Build Back Fairer.

Actually, it seems to me that Bodanis’s book is less about fairness than it is about the key word in the subtitle, decency. When you read the book, it often seems that the words decency and decent appear more often than fair and fairness, and certainly when you hear Bodanis speak about the book it is decency that seems to keep coming up. It seems an old-fashioned term, but it is none the worse for that; Bodanis is perhaps challenging us to refashion it for our modern era. That would be no bad thing, and I think I would agree that decency at least is an art.

The Art of Fairness, David Bodanis, Bridge Street Press 2020

Steve Ballmer: I love this company

Some thoughts on Rethink: build back fairer

Fairness, actually, is all around it seems this year. I know that my RAS (reticular activating system) is particularly alert to the term, but it isn’t just that: it’s clear that the Covid-19 crisis has brought many to the conclusion that fairness matters, and that focusing on fairness is more worthwhile than on the narrower numerical concern of inequality.

A key example is the apparent rebranding of the Rethink series on Radio 4 – which started with the Pope’s comments on poverty – to be Rethink Fairness. As host Amol Rajan explained, fairness was a term that just kept on coming up in the Rethink commentaries. This led to a series of five special discussions over last week, considering Rethink Fairness on: wealth, regions, education, health and generations.

The most striking part of the series was that both the education and health discussions focused less on disparities in education and health outcomes in themselves, and much more on those disparities as symptoms of broader unfairness in society. In each case, the suggested responses seemed more to do with addressing poverty and a lack of opportunity, particularly in some regions of the country, than with issues in education and health specifically. Both the health and education systems, it was noted, are too often asked to backfill for other failings.

Typical was this exhortation from the excellent Sammy Wright, vice principal of Southmoor Academy in Sunderland, and a member of Social Mobility Commission:

“We have to support people outside school, we have to have an adequate welfare safety net, we have to have services and we have to have people who can provide that whole wrap-around culture for struggling families.”

Similarly, “Good health is not about the health service, the health service is sorting that out when it’s gone wrong,” said the wonderful Dame Julie Moore, former nurse and recently retired chief executive of University Hospitals Birmingham NHS Foundation Trust. “The health service ends up sometimes picking up the price for the basic inequality in income, in housing, in circumstances.” She went further: “Sometimes the health service is there to fulfil a need that should not have arisen; we can’t expect the health service to solve all these problems.”

“Most of the difference that we see between children can actually be explained by their family background, their family income, factors outside of their school,” agreed Anna Vignoles, professor of education at the University of Cambridge. “It’s not surprising when you think about the influence of families and the amount of resource that they have, that they are making more of a difference than schools.”

Professor Sir Michael Marmot identified four mechanisms that in his view link our poor health pre-pandemic with our poor health during the pandemic. These four mechanisms are explored in detail in Build Back Fairer, last month’s report he co-authored on lessons to be learned from Covid-19 and reflecting on his decade of recommendations to the government for reform. Again, these are issues that go much broader than health as such:

  1. Quality of governance and political culture 
  2. Widening social and economic inequalities
  3. Investment in public services
  4. Health had stopped improving prior to the pandemic, which increased the risk of lethality during the pandemic

Vignoles put the call for action in similarly broad terms: “What we’re really facing pre-Covid, post-Covid, during Covid, is economic inequality in our society that also is mirrored in our education system and so if we’re serious about changing that we do have to work on the economic inequality that sits outside of schools first, or at least at the same time, because without tackling that we’re always going to face a problem in our education system.”

Furthermore though, it was clear that, just as many of the problems faced by the health and education systems arise from broader societal fairness issues, it is also true that addressing some of the health and education challenges – and investing more smartly in them – gives an opportunity to address future unfairness and create a nation in which we would all prefer to live. Getting these issues right can help with society as a whole, not just education and health in themselves. With so much focus on infrastructure spending, a key question is what infrastructure will actually matter for the future we want. There was some cynicism as to whether this should be physical transport links, and the suggestion that rather, it should be something smarter to suit the 21st century.

At its heart was a call not to see infrastructure just as concrete and steel. “View health as an investment in the nation’s infrastructure,” said Moore. “It’s a necessary part of our infrastructure, and it’s a really valuable investment on every front.”

Actually, the clearest articulation of the need for skills and investment in people and fairness rather than hard infrastructure came in the programme on Regions: “It’s not about faster trains between cities in the North of England, which would be used by richer people, it’s about trying to give people further down the ladder the skills that they need to get on in the world of work, and then also make that place a more attractive place to come and do business as well,” explained Paul Swinney, director of policy and research for the Centre for Cities.

This argument for a heavy investment in skills and training – and scope for retraining for those whose former jobs are no longer as valued – becomes more powerful when one considers the arguments of Harvard Professor Robert Putnam, of Bowling Alone fame. Putnam suggests that the rise of American public schooling in the progressive era (the years around 1900), which essentially provided a basic education for the entire population for the first time, readied the skills of its people and helped pump-prime the economy, providing the foundation for the economic success of the American century. 

If we aim to have a successful next century, we need to prepare our young people to succeed in it.

There were six areas of recommendations in the original Marmot Review, Fair Society, Healthy Lives, which again go far beyond a narrow understanding of health. They are a good basic set of ambitions for a nation to set itself:

  • Give every child the best start in life.
  • Enable all children, young people and adults to maximise their capabilities and have control over their lives.
  • Create fair employment and good work for all.
  • Ensure a healthy standard of living for all.
  • Create and develop healthy and sustainable places and communities.
  • Strengthen the role and impact of ill health prevention.

There seems no better message on which to end.

Rethink Fairness, Radio 4

Social Mobility Commission

Build Back Fairer: The Covid-19 Marmot Review, Michael Marmot, Jessica Allen, Peter Goldblatt, Eleanor Herd, Joana Morrison. The Health Foundation, December 2020

The Marmot review 2020 – the government must go further. Royal College of Physicians, February 2020

Health Equity in England: The Marmot Review 10 Years On, Michael Marmot, Jessica Allen, Tammy Boyce, Peter Goldblatt, Joana Morrison. Institute of Health Equity, February 2020

Fair Society, Healthy Lives, Michael Marmot, Jessica Allen, Peter Goldblatt, Tammy Boyce, Di McNeish, Mike Grady, Ilaria Geddes. Department of Health, February 2010

Twelfth Night, or whatever you want, CEO

It is Twelfth Night, the last of the Christmas festivities and by tradition when the decorations should be taken down. But Britain’s CEOs may be tempted to leave their decorations in place and risk the threat of the goblins that apparently invites, because today is also the day when top executive pay exceeds that of the average worker — named High Pay Day by my thoughtful friends at the High Pay Centre as a way of drawing attention to these differentials in pay between CEOs and the average full-time employee.

The bad news for FTSE 100 CEOs is that High Pay moment is a full hour later than last year as it now takes a CEO 34 hours of work to surpass average annual pay rather than the 33 hours it took last year. It is not that CEO pay has reduced; it has stayed relatively unchanged. Rather, average worker pay (perhaps counter-intuitively in these times of furloughing and Covid-driven unemployment) has risen a little. This pay ratio is around 120:1. High Pay moment this year is at around 5.30, about the time this blog is being posted.

Companies are now required to disclose their own internal pay ratios, that of CEO to the company’s average-paid worker in the UK (note ‘average-paid worker’, not ‘average worker’; no worker is average), and also a comparison of CEO pay to the lower and upper quartile worker in terms of pay. These ratios were analysed in another recent publication by the High Pay Centre, at whose launch late last month I had the privilege of being invited to speak. 

The results are striking, though the headline averages are lower than the 120:1 marked today:

For the FTSE 100, the median CEO/median ratio is 73:1,
and the median CEO/lower quartile ratio is 109:1 

For the FTSE 350 as a whole, the median CEO/median employee pay ratio is 53:1,
and the median CEO/lower quartile employee ratio is 71:1

These averages mask some remarkable variation, with the highest ratio, 2605:1, being that at Ocado, whose Tim Steiner enjoyed (and one assumes continues to enjoy) an extraordinary one-off level of pay of £58 million. Another seven companies saw CEO:median paid employee ratios over 200:1 (JD Sports, Tesco, Watches of Switzerland, GVC Holdings, Morrisons, CRH and WH Smith). The variation is shown a little in this chart:

It is clear from this that the biggest driver of the largest ratios is less the CEO’s individual pay (setting aside the unusual case of Mr Steiner) and more the pay of the workforce. That is still starker when the ratios in comparison with lower quartile workers by pay are considered. At Ocado, this ratio is 2820:1, at BP it is 543:1 and at Tesco it is 355:1. There are some sectors that are disproportionately represented among those with high ratios and notably low levels of pay — particularly retail and hospitality [Note that while AB Foods is classified as a ‘consumer goods’ company, its lower quartile paid workers are likely to be employed at its Primark subsidiary]:

In 11 cases, the High Pay Centre notes, the revealed lower quartile thresholds are below what would be earned from a 35-hour week paid at the £9.30 real living wage as calculated by the Living Wage Foundation (note, this is not the statutory minimum wage – no allegation of illegality is being made). We should note that in part these sectors being notably lower paid is an output of long-standing sexism in pay: retail and hospitality have traditionally been seen predominantly as women’s work and so have never been paid as well as other sectors. That is a clear unfairness.

And it is important to note that, as low as the pay revealed by these lower quartile figures is, fully a quarter of the workforces of each company is paid less.

When we focus on the lowest paid, it’s important to note as the High Pay Centre does that this data includes only those people that companies actually employ. HPC argues that outsourced workers should be included in pay ratios; I’m not convinced that this is practical, but I think it is realistic to ask for greater clarity of disclosure in business model reporting of all the additional workforce on which the company is dependent that are not necessarily staff members. More general enhancements to workforce discussions in business model reporting would also be helpful because these pay ratio disclosures include only UK workers. For some companies that will cover the whole business, while for others it is only a very small portion. Business model reporting can be made much more transparent and informative, and these sorts of disclosures would help towards a better understanding of how each company sits within its broader economic environment.

The High Pay Centre also calls for a number of ways in which the accountability of companies, and particularly remuneration committees, for pay decisions — including outcomes like pay ratios — should be enhanced. Personally, I have long said that there should be greater formal accountability of remuneration committees to staff members, in particular a formal presentation annually by the remuneration chair to employee representatives. I think that could prove salutary, and be a strong route for understanding staff perspectives that could then be fed into discussions in the boardroom. It might also help alleviate some of the ongoing suspicions that CEOs and other executives have too much influence on their own pay.

CEOs may be tempted to leave their decorations up. I suspect the rest of us will be moved to take ours down.

See also: Resentment and rents: fairness in executive pay

Fairness in the pay ratio

High Pay Day 2021, High Pay Centre

Pay Ratios and the FTSE 350: An analysis of the first disclosures, High Pay Centre, December 2020

Living Wage Foundation

The failures of algorithmic fairness

Technology will not of itself generate fairness. Merely asserting that the independence and lack of human intervention in a process removes bias isn’t true — indeed it may require human intervention to overturn failures to be fair. After all, fairness requires effort. This is the lesson that ought to be being learned in the world of algorithms, but unfortunately too often the creators and users of algorithms seem to prefer to state that they will be fair without actually carrying out the work effort that will deliver fairness in practice.

No parent of secondary school age children in England needs to be told that algorithms are not inherently fair. The debacle this summer of the supposedly standardised grades for students leaving school, who had lost the opportunity actually to sit their final exams because of Covid-19, brought the potential for technological unfairness fully into view. 

The key political decision was that there should not be significant grade inflation arising from the use of teachers’ assessed grades. The way that was operationalised created the unfairness, particularly the view that results awarded should be in line with schools’ historic performance (startlingly, almost 40% of teacher assessed grades were downgraded). The unfairness arose despite regulator Ofqual’s explicit aim being to deliver fairness for students. For example, the regulator required less standardisation of smaller classes than larger ‘cohorts’, because adjustments in smaller classes might appear more arbitrary. But a key consequence of this was that fee-paying private schools (which market themselves in part based on offering more esoteric subjects and smaller class sizes) faced fewer downgrades than free state schools. That looked like unacceptable unfairness to many.

Furthermore, the tying of performance to historic grades anchored exceptional years, or improving schools, to their less successful heritage. Fundamentally in a system that has to be all about what the individual her- or him-self deserves, an automated process for determining performance is always going to create some degree of unfairness at a granular level — especially so when the algorithm directly limited the number of certain grades that could be awarded to a particular school, condemning some individuals to multiple downgrades from those their teachers predicted for them. 

The scale and number of these apparent individual unfairnesses in the end led to a government decision to abandon the algorithm and simply revert to those teacher-assessed grades. No doubt there is some unfairness in that, but at least it appears less systematic unfairness.

So even where the aim explicitly is fairness, that is not necessarily what will be delivered if the design of the algorithm is faulty and it is subject to inadequate testing.

The scope for unfairness is much greater in cases where fairness is not the aim being sought. And where that application is to areas as sensitive as law enforcement and incarceration (as it increasingly is) then unfairness creates very fundamental problems. 

Take facial recognition, a technology increasingly used particularly at border crossings to identify those who can be allowed into a country and those who should be excluded. In theory perhaps the use of technology should remove bias. In practice, however, the technology is racist.

A remarkable December 2019 study by the US’s National Institute of Standards and Technology of more than 100 leading facial recognition systems, including ones from the most famous names in technology, reveals that even the best facial recognition systems misidentify black people at rates 5-10 times higher than the misidentification of white people. This chart is but one sample of the striking data the analysis reveals:

Differential success rates by nationality and gender, from one of the better recognition algorithms.
Source: National Institute of Standards and Technology

As is obvious from this chart, not only are the facial recognition systems racist, they are also sexist. They are not as sexist as they are racist, but the algorithms are consistently worse at identifying women correctly. They also happen to be worse at identifying older people and children, getting worse and worse at the increasing extremes of age. 

There is a consistent reason for these general errors: the development of the AI has been focused on one demographic, trained on a set of data. In most cases, that has predominantly been adult men from the locality of the developer. There is a strong home bias in the datasets used to develop the algorithms, meaning that there is a contrast between the data from most of the systems with the results from those developed in China: “with a number of algorithms developed in China this effect is reversed, with low false positive rates on East Asian faces”.

This reveals a general challenge about AI. There is a sense in which the creation of algorithms is based on the creators’ existing understanding of the world around them. Given that our world is riddled with unfairness, it is not surprising that unfairness is an outcome of their work — unless they deliver actions that actively work to remove the unfairness. 

As Hetan Shah, then Executive Director of the Royal Statistical Society, now CEO of the British Academy (and also vice-chair of the Ada Lovelace Institute and chair of the Friends Provident Foundation), sets out in an excellent brief paper on Algorithmic Accountability

“Algorithms for the most part are reflecting back the bias in our own world. A large part of ‘bias’ in algorithms comes from the data they are trained upon.”

It therefore follows that we need to lean actively against existing bias and prejudice in order to make algorithms more fair. Among the steps which would assist in this, Shah argues, are enhancing the diversity of the technology workforce, ethics training and a professional code of conduct for all data scientists, and new deliberative bodies to help set standards, and build ethics capacity.

The tech industry is alive to at least some of these issues — or sounds like it is. The language of fairness is used freely in its discussions of its own work. For example, in January 2018 Microsoft published its “ethical principles” for AI, which start with ‘fairness’. In May of that same year, Facebook announced its “commitment to the ethical development and deployment of AI” which includes a tool that it claims can systematically seek out and identify bias in algorithms, called ‘Fairness Flow’. That September saw IBM announce its ‘AI Fairness 360’, which is similarly designed to “check for unwanted bias”. The word appears prominently among the stated intents of Alphabet (Google) and Amazon.

The firms have also freely sponsored academic and other programmes supporting the development of fairness in AI, such as the Algorithmic Fairness and Opacity Group at Berkeley, that same University’s Center for Technology, Society & Policy (which has a project on Just Algorithms: Fairness, Transparency, and Justice), and the US National Science Foundation Program on Fairness in Artificial Intelligence.

Rodrigo Ochigame, who has worked within at least one of these organisations, is cynical about this: “Today’s champions of ‘algorithmic fairness’, sometimes sponsored by Silicon Valley firms, tend to frame discrimination and injustice as reducible to the stated distinction between the optimization of utility and other mathematical criteria”. His charming short article on the history of algorithmic fairness travels far — from 17th century English real property law, through 19th century US insurance contracts — and makes it clear that delivering fairness is not, and never has been, as simple as the champions of technology want it to seem.

In an earlier, excoriating article, Ochigame argues that the current efforts are little more than active lobbying for limited regulation. “Silicon Valley’s vigorous promotion of ‘ethical AI’ has constituted a strategic lobbying effort, one that has enrolled academia to legitimize itself,” he argues. This lobbying effort “was aligned strategically with a Silicon Valley effort seeking to avoid legally enforceable restrictions of controversial technologies”. 

Certainly, despite all the language of fairness, the example of the facial recognition algorithms suggests that there is still a long way to go to deliver fairness in practice. Ochigame identifies a number of other uses of algorithms in law enforcement and the criminal justice system that appear in practice to be delivering clear unfairness. 

The US’s Brookings Institute shares these concerns, highlighting a series of biased outcomes from algorithms, including gender bias in an Amazon recruitment algorithm, and bias in the online advertisements shown to online searchers, reinforcing existing prejudice and closing the door to difference. Just as with the facial recognition algorithms, the approach of building from what already exists or is near at hand leads to algorithms “replicating and even amplifying human biases, particularly those affecting protected groups”. This tendency, Brookings argues, is particularly concerning where technology is used in the criminal justice environment, to identify ‘potential’ criminals or sites of crime, or to help determine sentence lengths or the availability of bail. The solution, they believe, is not a technology one:

“companies and other operators of algorithms must be aware that there is no simple metric to measure fairness that a software engineer can apply, especially in the design of algorithms and the determination of the appropriate trade-offs between accuracy and fairness. Fairness is a human, not a mathematical, determination, grounded in shared ethical beliefs. Thus, algorithmic decisions that may have a serious consequence for people will require human involvement.”

Indeed, it seems as though algorithms risk becoming in some way a technological version of the sus law, the 1970s use of long-standing laws enabling the UK police to stop and question any individual suspected of intent to commit an arrestable offence. The sus law became a vehicle for blatant prejudice and harassment of ethnic minorities, and was repealed in 1981 following a series of race riots. The crudeness of social profiling by the police in the 1970s — even senior officers claiming that disproportionate targeting of black people was justified because they were over-represented among robbery and violence offenders — is reflected in much of the algorithmic modelling now being applied. Using technology does not make the work more fair or justifiable. It is just prejudice in another, perhaps less accountable, form.

The use of algorithms in the business arena has also been found to introduce unfairness. A recent study considers the German retail petrol (gasoline) market, which started to adopt algorithmic pricing in a significant way in 2017. Essentially, this study provides evidence that such pricing algorithms can drive anti-competitive behaviour: adoption of algorithms increases margins by 9% on average, even though non-monopoly markets show no margin enhancement at all. Looking just at duopoly markets, the researchers conclude: “we find that market-level margins do not change when only one of the two stations adopts, but increase by 28% in markets where both do.” The consequence is “a significant effect on competition”. It should probably come as no surprise that pricing algorithms push prices up; the question is whether this can be justified or whether it just amounts to a technological cover placed over the gouging of customers.

So fairness continues to be a victim in practice of the use of at least some algorithms, despite the apparent efforts by the technology giants to promote and assure the delivery of fairness. And rather than coming from the organisations they fund, in practice the best outlines of what is really necessary to deliver fairness in artificial intelligence — at least that I have been able to identify — come from other sources.

Of course the EU’s High-Level Expert Group on AI includes individuals from many of the leading technology firms, but the group is wider than that. In April 2019, it presented its Ethics Guidelines for Trustworthy Artificial Intelligence. Among the seven key requirements that AI systems need to meet to be deemed trustworthy, is the fifth, Diversity, non-discrimination and fairness. This states: “Unfair bias must be avoided, as it could could have multiple negative implications, from the marginalization of vulnerable groups, to the exacerbation of prejudice and discrimination. Fostering diversity, AI systems should be accessible to all, regardless of any disability, and involve relevant stakeholders throughout their entire life circle.”

In July 2020, the High-Level Expert Group took this work further and presented its final Assessment List for Trustworthy Artificial Intelligence. This provides questions for assessing whether AI delivers on the trustworthiness measure that the Group set. Among the questions under Diversity, non-discrimination and fairness are:

Is your definition of fairness commonly used and implemented in any phase of the process of setting up the AI system? 

  • Did you consider other definitions of fairness before choosing this one? 
  • Did you consult with the impacted communities about the correct definition of fairness, i.e. representatives of elderly persons or persons with disabilities? 
  • Did you ensure a quantitative analysis or metrics to measure and test the applied definition of fairness? 
  • Did you establish mechanisms to ensure fairness in your AI system?

This looks very similar to the simple model that Hetan Shah suggests for reducing and perhaps removing bias: (1) pilot to check for bias in multiple ways, with different datasets; (2) offer transparency to enable external scrutiny; (3) monitor outcomes for differential impacts; (4) provide a right to challenge and seek redress; and (5) enable enhancement through good governance (e.g. through independent oversight). In the context he is considering, of public data being used in algorithms by private companies, he also suggests (6) use of negotiating strength by the public sector as monopoly owner of data which private sector rivals are competing for.

Brookings reaches similar conclusions, proposing the idea of ‘Algorithmic Hygiene’, “which identifies some specific causes of biases and employs best practices to identify and mitigate them”. It is unsurprising therefore that one of their key recommendations for enhancing fairness is to: “Increase human involvement in the design and monitoring of algorithms”.

As an aside, this danger that arises from assuming our starting point needs to be the world as it is and as we understand it is consistent with other concerns — as Brookings notes, the biases in search arise from similar errors. I worry about algorithms constraining our horizons and feeding confirmation biases. In particular, I worry about the role of algorithms in defining search and so much of our online lives, of cookies constraining what we see. This removes the joy of serendipity and happenstance. A friend complains that using any search engine other than his usual one of choice (you can guess which) means that he sees all sorts of untargeted, irrelevant material. Perhaps his approach is more efficient by fractions of seconds, but inefficiency occasionally has its benefits and I suspect we lose much by seeking to avoid it — or by handing the power of happenstance over to a machine which tends to want to confirm our certainties rather than enable us to happen upon challenge and difference. The algorithm is the echo chamber.

Algorithms, like any human technology, are neither fair nor unfair. They are not automatically fair, as their IT proponents would like us to believe; nor are they automatically unfair as many campaigners seem ready to argue. Like any human technology, they reflect the prejudices and understandings of their creators and the society in which they are created. We live in an unfair world and so most technology, if it does not actively lean against unfairness, will be unfair. There needs to be much more work to ensure that algorithms operate fairly — merely being technology does not deliver that, without much, much more. A technology industry that fails fully to engage with this challenge will not build the necessary trust and will see confidence in algorithmic technology erode. Current unfairnesses suggest that the industry has much more work to do.

Small corrections made October 18 2022 following input from a member of Data Ethics Club. With thanks.

See also: Learning from the Stochastic Parrots

Awarding GCSE, AS, A level, advanced extension awards and extended project qualifications in summer 2020: interim report, Ofqual, August 2020

Face Recognition Vendor Test (FRVT) Part 3: Demographic Effects, Patrick Grother, Mei Ngan, Kayee Hanaoka, National Institute of Standards and Technology, December 2019

Ada Lovelace Institute

Algorithmic accountability, Hetan Shah, Philosophical Transactions, Royal Society, A 376: 20170362, April 2018

The Long History of Algorithmic Fairness, Rodrigo Ochigame, Phenomenal World, January 30 2020

The Invention of ‘Ethical AI’ – How Big Tech Manipulates Academia to Avoid Regulation, Rodrigo Ochigame, The Intercept, December 20 2019

Algorithmic bias detection and mitigation: Best practices and policies to reduce consumer harms, Nicol Turner Lee, Paul Resnick, and Genie Barton, Brookings Institute, May 22, 2019

Algorithmic Pricing and Competition: Empirical Evidence from the German Retail Gasoline Market, Stephanie Assad, Robert Clark, Daniel Ershov, Lei Xu, CESifo Working Paper No. 8521, August 2020

Ethics Guidelines for Trustworthy AI, High-Level Expert Group on AI, European Commission, April 2019

Assessment List for Trustworthy Artificial Intelligence, High-Level Expert Group on AI, European Commission, July 2020

Data Ethics Club

Fairness – the human lens for addressing our current challenges

A version of my presentation to the Church Investors Group (CIG), November 20 2020

My name is Paul Lee, and I am an investor. I have been in the City for 20 years now, first for more than a dozen years at Hermes, helping create and build its Equity Ownership Service (EOS) business and most recently at Aberdeen Asset Management, leading its stewardship efforts, and leaving following the merger with Standard Life. I am now an independent consultant, mostly to investment institutions.

But I am here to talk with you today about fairness. It’s an issue that I have done a good deal of thinking about over the years, and most particularly over the last two years in which I have been writing a blog on the topic. That blog and the issues it discusses seem to have gained a good deal more resonance in this year of Covid when many more people have been thinking about the role of business in society and how business, and investment, can more clearly be seen as working in the interests of all of their stakeholders.

I’m going to talk about two things in particular:

  • First, how fairness is central to how we as humans view the world, and central to our understanding of how we want our world and our communities to be.
  • Second, how this understanding might help us think about our role as investors and how we might best address some of the challenges our world faces.

Before that though, I should just talk about what fairness is. It is actually one of those things that is hard to define, but we certainly, viscerally know when it is absent. Unfairness is something that troubles people on a very fundamental level and language around fairness is central to how we want our world to be: we talk about level playing fields and removing unfair barriers to both competition and individual advancement. But fairness is not the same as equality; it is a sense, a feeling, not a number. 

Economists love inequality. Actually, some economists probably do love inequality but what I mean to say is that economists love the concept of inequality and talking about it. They measure it in precise detail, usually using the Gini coefficient, which measures disparity in either wealth or income, and largely ranges in practice from around 0.65 for the highly unequal South Africa to just under 0.3 for more egalitarian Scandinavian countries. For me, the fact that inequality can be reduced to single numbers like this makes it less meaningful for people. 

But crucially, inequality is less interesting than fairness because humans can stand some inequality – it is acceptable for us where individuals do better because they have particularly valued talents, or are particularly hard-working, or even where they enjoy particular luck. What is not acceptable, the inequality that humans cannot stand, is where that inequality arises from unfairness. And I don’t think it is controversial to say that this is something that the world is currently suffering from.

Now I promised you a human view of fairness. But actually I am going to start with some of our brethren, capuchin monkeys. They also display a sense of fairness, as can be seen in this video. It is taken from a talk by a great Dutch zoologist Frans de Waal who has done more than anyone to reveal the inner lives of our ape brethren (I would heartily recommend his book Mama’s Last Hug in particular). These capuchins are happy to perform a small task for the reward of a piece of cucumber. That is, they are until it seems that their fellows are rewarded for the same task with a grape. Let’s see.

From Frans de Waal’s TEDTalk 2012

So capuchins have a sense of fairness. But it seems to me a selfish form of the sense of fairness, righteous anger when another does better out of a given situation. It’s the same sense of fairness expressed by the six-year old stamping their foot and yelling “It’s not fair” when things do not turn out precisely to their liking. Yet humans show not just this selfish form of the sense of fairness but an unselfish form, an altruistic sense of fairness. Indeed, there’s evidence that even 15 month-old humans can show an altruistic sense of fairness.

Adult humans show this altruistic sense of fairness too – and apparently universally so, across cultures, across geographies. The main way that this has been studied and shown by social scientists is through two tests, called the Ultimatum Game and the Dictator Game

In the Ultimatum Game an individual is given a pot of money and invited to give any chosen portion of it to the other player. The recipient has a simple choice – either they take what they are offered and the two players walk away with their resulting outcome, or they refuse, in which case neither party gets anything at all. Now, if we listened to economists and were actually their version of humanity, homo economicus, the recipient would accept any offer that left them with something other than zero. But we aren’t, and I am sure you are all imagining what you yourselves might do in this game, what offer you might make and what offer you would be willing to accept. We’ll look at some real world evidence in a moment.

The Dictator Game is a simplified version of the Ultimatum Game. Again the offeror is given a pot of money, but this time they are a dictator because whatever allocation they make automatically happens. The recipient is given no choice to make at all. Clearly, in this game only one of the two individual’s view of the sense of fairness matters rather than it being a matter of the interaction of two individuals’ sense of fairness. Again, no doubt you are considering what outcome you might impose were you a dictator.

There have been many studies, and here is just an example of the largely consistent outcomes from them: 

From Markets, Religion, Community Size, and the Evolution of Fairness and Punishment, Henrich et al, Science Vol 327 (2010)

This study considers a range of scales and styles of community and tries to identify differences. The first thing to note though is the consistency– nowhere in this study does the average dictator proffer less than a quarter of the pot, even though economics would argue that they should hand over nothing. In many places the award is near to half of the overall pot – even in the highly unequal and unfair US (about which more later). Again, in most communities of any size the study shows that really the only acceptable Ultimatum Game offer is around 50% of the pot. 

One might suggest that the reason why smaller communities, or those less involved in markets, are less demanding of fairness in each transaction is that they have a greater sense that what goes around comes around – that there will be a further transaction with the individuals in question and fairness can be achieved over time. In larger communities (anything more than a small village to most of our perceptions) or where commercial life operates, it seems less certain that there will be further interactions so perhaps we feel obliged to seek fairness at each opportunity.

Members of this audience in particular might have noticed the word ‘religion’ in the title of this study. And you may be gratified to hear that it provides some evidence that participation in ‘world religions’ (here defined simply as either Islam or Christianity) is associated with a greater sense of fairness. However, I am not sure that it is very strong evidence as the only studied community with zero participation in world religion is the Hadza, one of those in the bottom left corner of the dictator game chart; all the others have well over 50% participation, and most are at or near to 100%, so it’s far from clear that that particular finding is strongly meaningful.

While we are on the topic of small communities I want to talk a little about the Ju’hoansi tribe in Namibia. This is a hunter-gatherer tribe, and believed to be one of the few to have a culture that has been consistent for tens of thousands of years. It may therefore offer the best window we have to the cultural outlook that framed the development of the human brain. And the Ju’hoansi have a very interesting approach to success. A hunter returning with a kill is not celebrated but rather teased. The kill itself is denigrated – it was barely worth bringing back to the camp, it will barely feed anyone – that sort of thing. There is a strong sense that individual success should not be praised but rather should be brought low, because what matters is success of the community as a whole – in a real sense, this can be a matter of life and death.

It is interesting to contrast this Ju’hoansi mindset with our own. We believe in merit, we believe in rewards for success. We like to believe that we live in meritocracies – in fact, believing that we do reflects our need for fairness – and so rewarding success is only right. We can perhaps see echoes of the Ju’hoansi approach in the banter that we experience in office environments, and in the tall poppy syndrome to which we are prone, where we tend to denigrate those among us who stand out. But we also see the opposite:

From Executive pay: review of FTSE 100 executive pay, CIPD and High Pay Centre, August 2018

This is a chart from my friends at the High Pay Centre. They are currently updating the numbers to the latest year, but from what I have seen the shape is not markedly different. It shows in effect the opposite of the Ju’hoansi mindset: it shows substantial reward, and in effect praise, for the successful. Just to be clear about the numbers – the red bars show CEO annual pay in thousands, so ranging from over £9 million in the oil and gas business down to a ‘mere’ £2.5 million in utilities. The line shows the ratio between this and the reward for the average worker – sorry, there is no such thing as an average worker – the worker paid in the middle of the range. This ratio shows the variation of CEO pay being around 50 times average pay, to 85 times, 140 times, or indeed 277 times.

I’m sure that there is nothing too surprising in this chart, but I will draw your attention to two things. First, the right-hand side, the two most generous paying sectors – oil and gas and consumer goods. It seems relevant to note that in those sectors the UK has companies that genuinely are global competitors (BP and Shell, and Unilever and Reckitt Benckiser or RB, respectively) and so compete with US rivals; there is a marked contrast in the pay levels with, say the utilities sector, which is a UK domestic industry. The second thing to note is that the ratio is driven less by the pay of the CEO than by the nature and structure of the businesses in question. For example, consumer services, which has a high spike in the ratio, includes retail where we know there are many part-time workers and many who are low-paid; in contrast, the financial sector ratio is much lower because as we know many in financial services are tremendously well-paid.

No doubt all those who are well paid believe that they are well worth it. We live in meritocracies and so of course those who succeed deserve their success. That is the myth of meritocracy, which those who have some success are keen to believe in. They tend to emphasise their own merit and their own hard work, and perhaps to ignore the element of luck that no doubt helped them towards their success – perhaps luck in their start in life or in terms of a lucky break, or otherwise. We do not live in as meritocratic societies as we might like to believe, but we still fall prey to the myth of meritocracy. And the evidence is that believing in meritocracy allows us to put up with more inequality than we might otherwise do: people around the world are much less likely to correct manifestly unequal allocations if they are told there is some (however tenuous) connection to merit or hard work than they are if they are told it arises from luck.

And while we believe in meritocracy, we do not live in meritocracies: 

From Perceptions of Inequality Using Social Economics Surveys, Stefanie Stantcheva (Harvard University) January 2020

These charts are from a recent presentation by a US academic, looking at perceptions of how meritocratic societies are – which, after all, is one measure of how fair they are. This shows how difficult it is for someone born into the bottom quintile of society financially – the bottom 20% – to escape from that quintile, and also to break into the top quintile – the top 20% – financially. Just to state the obvious, in genuinely meritocratic societies these numbers would be 20%, or very near to it. And all of the countries that we are looking at here are a long way from that 20%. But it’s worse for the US, because they are frankly delusional about how meritocratic their society is. In both senses US citizens believe their country is more meritocratic than the perceptions in the other countries studied, and in both cases it is by some margin the worst in reality. This enables that nation to accept much more inequality than other countries – truly that fantasy is the American Dream.

While we may enjoy a little more realism in the other countries, we cannot pretend to ourselves that they are genuinely meritocratic as each is a very long way from 20% in both cases. Perhaps we need a little more Ju’hoansi thinking.

So that is a very brief view of what fairness is, how central it is to our view of ourselves as humans and of the communities in which we want to live, and of some of the implications of the distance we currently are from fairness. 

So to turn to what this might mean for us as investors. I have one suggestion for a space to explore in terms of investment opportunity, and three areas (each with two arms) to consider in our broader investment and stewardship approach.

Taking the investment opportunity first, I believe that there must be an opportunity in providing opportunity, in leaning against the unfairness and barriers to advancement in our society. That might mean investing in deprived areas to generate economic activity and more opportunity for those who currently face barriers. Given that successful investing is buying cheap assets in the expectation of valuation improvements, at the very least it is worth considering as there are cheap assets to invest in and improve.

To take the other three areas in turn:

  1. Supply chains
    1. First, there is an economic problem with supplier relationships. The purchasers, typically larger companies, require suppliers to finance the supplies – they take goods on credit, at 30-day terms or 60-day terms. So the cost of financing the supply chain sits with the supplier rather than the purchaser. Given the purchaser is as I say typically larger, its cost of finance is lower so there is a clear economic cost from this unequal relationship.
    2. Second, the impact of the downwards pressure on pricing leads to bad things happening in supply chains. We have seen stories just this week about inappropriate behaviours at suppliers in India, we know the stories about the abuses within fast fashion supply chains in the great city of Leicester; more broadly, we know deforestation is driven by our demand for cheap products. We need to look harder at supply chains.
  2. Tax
    1. The responsible tax movement is growing and it is really important. More companies need to be challenged to be responsible taxpayers, to ensure they pay a fair level of taxation in their countries of operation. Too much profit is syphoned away into tax havens and we need more honesty and fairness over tax.
    2. We as investors need to look at our asset classes beyond listed equities. In particular, we should look at our private equity and fixed income portfolios. Too often, companies are thinly capitalised and financed with heavy debt loads, reducing their profits and so their tax load. Too much of private equity return arises from this tax arbitrage, and possibly too much of the debt in many portfolios is playing the other part in this game. Companies with too much debt have proven not resilient enough in this crisis, and we need to consider this actively – as well as the simple fairness of ensuring that the exchequer receives a fair level of tax.
  3. Executive pay
    1. I would actually suggest that in the UK and Europe it would be better to look not at CEO pay – remember that in the UK we have successfully held down CEO pay for around the last seven or eight years – but at the other end of the pay ratio, the pay of the least well off. I would suggest a focus there, on contractors as well as staff, might lead us nearer to fairness.
    2. Put simply, if you are not already leaning hard against the levels of executive pay in the US, you should be. As we saw in the rewards for consumer goods companies and oil and gas businesses, the high levels of executive pay in the US infect the world.

Thank you for the opportunity to talk a little about the sense of fairness and also what it might mean for us as investors. I’d welcome thoughts and questions.

Mama’s Last Hug, Frans de Waal, Granta (2019)

Fairness Expectations and Altruistic Sharing in 15-Month-Old Human Infants, Marco Schmidt, Jessica Sommerville, Plos One (2011)

Markets, Religion, Community Size, and the Evolution of Fairness and Punishment, Henrich et al, Science Vol 327 (2010)

Envy’s hidden hand, James Suzman, Aeon, May 2018

Executive pay: review of FTSE 100 executive pay, CIPD and High Pay Centre, August 2018

Perceptions of Inequality Using Social Economics Surveys, Stefanie Stantcheva (Harvard University) January 2020

We will be judged: the knell of intergenerational fairness

“For me the question now after six months of the outbreak, the question remains to the elders and decision-makers, what kind of world are you leaving for us? Is it fair that we inherit this unequal, violent, empty of values world? We are paying for a system we did not co-design and yet we are inheriting this anyway.”

So said Aya Chebbi, the African Union’s first Youth Envoy, at the Tortoise G7bn Summit in September. Where youth has a voice, it is already calling older people to account. If we listen, we are being reminded that we have not been, and are not being, good ancestors.

The challenge of intergenerational fairness is one that our world is currently failing. A recent WHO-UNICEF-Lancet report, A Future for the World’s Children?, found that current policies are failing future generations in every country in the world. Bluntly, it reports that in spite of the dramatic improvements in survival, education, and nutrition for children worldwide over the last 50 years, meaning that “In many ways, now is the best time for children to be alive”, nonetheless “economic inequalities mean benefits are not shared by all, and all children face an uncertain future. Climate disruption is creating extreme risks from rising sea levels, extreme weather events, water and food insecurity, heat stress, emerging infectious diseases, and large-scale population migration. Rising inequalities and environmental crises threaten political stability and risk international conflict over access to resources. By 2030, 2.3 billion people are projected to live in fragile or conflict-affected contexts.”

One chart from the report is particularly stark, showing a clear correlation between income inequality and measures of child flourishing (data newly synthesised for the report), particularly relative to countries of a like income bracket. “Equity [fairness] is essential to ensure that efforts to promote children’s present and future flourishing truly leave no one behind,” it notes. There is an implication for global fairness from this analysis: if global inequity worsens, so will the prospects for our children; conversely if we can improve fairness we have the opportunity to enhance the prosperity and wellbeing of all.

The WHO-UNICEF-Lancet team suggest that children’s interests should be placed at the heart of the Sustainable Development Goals (SDGs), the UN’s ragbag charter of 17 global development aims for the 15 years from 2015, from zero hunger and clean energy to climate action and responsible consumption. “Fundamentally, the SDGs are about the legacy we bequeath to today’s children. For that reason alone, children should be placed at the centre of the SDG endeavour,” the report suggests. It also references the General Comments to the Convention on the Rights of the Child, which includes the right to “Be treated fairly”. Our record against the SDGs suggests we are not succeeding in fair treatment.

But this is not a challenge for developing economies only, and not a challenge that only developing economies are failing. Intergenerational unfairnesses are stark in many developed economies. Take the Resolution Foundation’s Intergenerational Centre statistics for the UK, as exemplified in the following chart from their excellent interactive data dashboard tool. This shows relative poverty after housing cost over time for the population as a whole (olive line), the over 65s (sage line), the under 17s (pink line) and those from 16-29 (purple line):

While the headline poverty number for the population has risen somewhat, from around 13% in 1961 to just under 20% in 2017, this masks radically different experiences for different portions of the population. The pension aged population has gone from a scandalous 40% living in poverty to around 15%, while the proportion of children living in poverty has nearly trebled from just over 10% to just under 30%. But the trajectory of the 16-29 age group is perhaps most startling: having seen by far the least poverty in 1961, at only 4%, they are now well above the population average at 22% in 2017 (and 27% as recently as 2012). 

As Aya Chebbi’s comment indicates, there is a short-term, virus-related element of these unfairnesses, and also a longer-term aspect, particularly in relation to the scale of the climate crisis. Taking the immediate issue first, it is clear that Covid-19 will further erode fairness between generations. As can be seen from this chart from Resolution Foundation’s recent intergenerational audit, there is a stark age effect in furloughing, and still more so in jobseeking. And it is set to worsen further: Office for Budget Responsibility estimates see unemployment for the 18-29 age group spiking to 17%; older groups are expected to see unemployment no worse than 7% (though of course that is bad enough). While these statistics are for the UK, the experience is likely to be universal given the jobs usually taken by the young — both less stable and skewed towards sectors hardest hit by the virus and the constraints it places on our lives. Some of this is inevitable, but it could be mitigated better.

We are making life harder for the generations that come after us. That’s the opposite of what we are called to do.

In his reactionary 1790 diatribe against the French Revolution, Edmund Burke chastised those who do not recognise that “we are temporary possessors and life renters” of our world and our society. He criticised those who are “unmindful of what they have received from their ancestors, or of what is due to their posterity” and instead “act as if they were the entire masters”. He charged that we “should not think it amongst [our] rights to cut off the entail, or commit waste on the inheritance”. 

While that was a highly conservative message, it is clear that refusing to change can in some circumstances cause just as much damage to future generations as can revolutionary change. 

It is the very heart of good stewardship to build in thoughts about the future. As temporary possessors, we are all called to operate with consciousness of the needs of the future rather than assuming it will be able to look after itself. Life renters must find a way to resolve the tension between short-term pressures and longer-term needs. For example, Edward Laurence in his 1727 The Duty of a Steward to His Lord (which I elsewhere identified as The Original Stewardship Code) discouraged the use of clay for bricks as that depletes the land; similarly, he charged that stewards should not sell off timber that rather needed be retained for use as building material within the estate. His tome is almost obsessed by maximising the availability of fertiliser to enrich the estate’s soil for the future.

However, instead of acting as good stewards in this way, we continue to deplete the world’s resources on a Micawberish assumption that something will turn up.

More recent thinkers have reached similar conclusions. In 1977 economist John Hartwick set out what has become known as the Hartwick Rule of intergenerational equity. He stated that “to ensure intergenerational equity, society should invest enough of the rental income from extraction and use of exhaustible, and thus naturally scarce, resources so that future generations would benefit as much as today’s”. As the World Bank puts it in The Changing Wealth of Nations: Measuring Sustainable Development in the New Millennium (2011): “The Hartwick rule holds that consumption can be maintained — the definition of sustainable development — if the rents from non-renewable resources are continuously invested rather than used for consumption.”

John Rawls, the great thinker on fairness whose ideas have become decidedly unfashionable as the liberal American postwar economic success that he sought to justify has receded from view, posited a similar just savings principle. This argues that there is a duty to save justly such that future generations enjoy at least the minimal conditions necessary for a well-ordered society. In other words, we must not consume so much now that it is at the expense of future generations, instead we should set aside savings to benefit future generations. This goes one stage beyond Hartwick because it contemplates not only avoiding a negative such as limiting depletion of scarce resources but also actively pursuing positive measures, including investment in education and technology.

The perceived limited resources have changed over time. Once it was wood or mineral wealth. Now the most urgent limitation is seen to be the capacity of our atmosphere to absorb further CO2 and other greenhouse gases without further increases in global temperatures. There is some investment in technology to mitigate carbon intensity, but many doubt it is anywhere near sufficient. But the philosophical mindset remains the same: we have a duty to our successors not to bequeath them a worse world. We are failing in that duty currently.

Some are at least considering how the interests of future generations can be more fully built into policy-making and current planning. “The fear is that today’s adults are mortgaging our children’s future; taking too many policy decisions with an eye to short-term benefits while disregarding the long-term harms. In the years ahead, the sense of intergenerational tension is set to intensify, with more generations alive simultaneously and tightening availability of limited resources, from ecological resources such as water and forests to pensions savings,” say Cat Tully, the School of International Futures and Luis Xavier, Calouste Gulbenkian Foundation, in a June 2020 article on designing policies that are fair to future generations. 

They highlight the 4 key principles and approaches to intergenerationally fair policymaking identified by the Gulbenkian Foundation’s Intergenerational Fairness initiative. These offer a method for discerning policies that do deliver fairly between generations. They are written for the world of public policy, but applicable with some sensible adjustment to help shape corporate strategic thinking (essentially, all are about giving space for the interests of neglected stakeholders to be considered):

  • Run regular “national conversations” to (a) engage the public in exploring possible futures for their country, and (b) devise the detailed measures of intergenerational fairness for the framework. 
  • Give immediate responsibility for “vetting” policy for its fairness or unfairness to an independent government body or bodies.
  • Put in place the wider conditions to lock in the institutional and public pressure for the new framework to work.
  • Activate “outside-in” (public-on-government) pressure.

Some business leaders are already trying to give effect to such a mindset. Dan Labbad at the British Academy’s Future of the Corporation Purpose Summit referred to future generations as “arguably the most important stakeholder” for his business. But, given he is CEO of the Crown Estate, which runs the assets of the British nation’s monarch, he has an unusually long time horizon for operations, not least as “We have a statutory obligation to protect, maintain and enhance the estate into perpetuity”. Nonetheless, he argues that future generations are an important stakeholder common to all businesses, “often neglected, but crucial”:

“While they don’t get to hold us to account today they are the recipients of what we leave behind. That’s why our purpose at the Crown Estate will be guided by the creation of something that ultimately represents them and their interests to ensure that they are more than mere recipients but ultimately beneficiaries, inheriting something that will be the foundation of their future, a legacy that they’ll be proud of. I strongly suggest we take their role as a stakeholder seriously and as a humble reminder of why we need to operate with purpose in the first place.”

Building the interests of future generations into our decisions now is something we seek to do (not always successfully) when we think about our own children, and grandchildren. But we struggle to do it on a grander scale, and yet we need to. Certainly, most business fails to live up to this demand. Instead, we are mortgaging future generation’s inheritance now, and that is not fair. We are bequeathing a worse world, not a better.

Where youth has a voice, it is holding older generations to account for their unfairnesses. We need to ensure that youth has a voice, and that their voice is heard, so that they can begin to help co-design their own future. As we have failed to do so in the past it is perhaps not surprising that we have caused so much damage to their future, notwithstanding the longstanding calls to operate with a long-term mindset that does provide them with a fair world to grow up into.

“Is it fair that we inherit this unequal, violent, empty of values world? We are paying for a system we did not co-design and yet we are inheriting this anyway.”

I am grateful to my friend Peter for challenging me to write about this topic finally, and flagging up the article on apolitical (sorry it’s taken a while!)

Tortoise G7bn Summit

The Good Ancestor, How to Think Long Term in a Short-Term World, Roman Krznaric, 2020

A future for the world’s children? WHO-UNICEF-Lancet, February 2020

Resolution Foundation Intergenerational Centre

An intergenerational audit for the UK, Resolution Foundation, October 2020

Reflections on the Revolution in France, Edmund Burke, 1790

The Duty of a Steward to His Lord, Edward Laurence, 1727

The Changing Wealth of Nations: Measuring Sustainable Development in the New Millennium, World Bank, 2011

How to design policies that are fair to future generations, Cat Tully, Luis Xavier, apolitical, June 2020

Gulbenkian Foundation Intergenerational Fairness initiative

What is the role of stakeholders in purposeful business?, from the Future of the Corporation — Purpose Summit, British Academy

What’s the Purpose of Purpose? (II)

It turns out that those of us who have been cynical about the Business Roundtable statement on the purpose of the corporation (see the Limited Responsibility Company and What’s the Purpose of Purpose?) were right. There is strong evidence that the statement is words merely, with no intent for action.

“we show that the statement is largely a rhetorical public relations move rather than the harbinger of meaningful change”

These are the conclusions of a forthcoming article in Cornell Law Review by the dazzling Harvard law professor Lucian Bebchuk and his colleague Roberto Tallarita, flagged through an op-ed in the Wall Street Journal earlier this month. They start with the assertion that the Business Roundtable statement is “remarkably vague as to the nature and content of the commitment that is being made” and conclude that there is no real underlying commitment at all.

The most concrete evidence of this that Bebchuk and Tallarita bring is to have asked the 173 companies who signed the BRT statement what was the highest authority in the organisation that approved the signing of the statement. Only 48 corporates replied, and of those 47 stated that the CEO had approved the signing without the involvement of the board. Only a single company said that it had taken the decision for board approval. Yet the academics note that this should be no personal assertion, but rather a top-level decision about how the company is to operate going forwards. The statement explicitly says that signatories “commit to lead their companies for the benefit of all stakeholders”, and it is expressed as a full redefinition of the corporation. 

In spite of these words, signatories seem to assume that nothing has changed: “two of the companies that responded to our survey stated that joining the BRT statement reflected an affirmation that the company’s past practices have been consistent with the principles of the BRT statement rather than an expectation that the company would make major changes in its future treatment of stakeholders”. Furthermore, a third company, JP Morgan, asserted that it has always operated in accord with the statement and will continue to do so.

So nothing has changed, and the statement does not have meaningful effect in spite of the fanfare with which it was introduced. This is reflected, Bebchuk and Tallarita note, in a lack of change to board corporate governance guidelines, which continue to focus exclusively on shareholder interests. It also reflects the ongoing focus on shareholder interests in executive pay. At least in the sample of signatories the academics considered, there is no reference to stakeholder interests as a driver of executive pay. “With strong incentives to care about shareholder value, and little incentive to care about stakeholder interests, CEOs are discouraged from making any decisions that would benefit or protect stakeholders beyond what would be necessary for shareholder value maximization.”

The lack of admission that there are trade-offs between stakeholder interests is a fundamental flaw in the statement, according to Bebchuk and Tallarita. They single out the wording that “while we acknowledge that different stakeholders may have competing interests in the short term, it is important to recognize that the interests of all stakeholders are inseparable in the long term.” Bluntly, Bebchuk and Tallarita say, “This is at best a naïve misunderstanding or, more realistically, a mischaracterization of economic reality.” In practice, they argue it will mean that shareholder value drivers will continue to predominate.

The academics argue that there needs to be clarity on how to approach the challenge of balancing different stakeholder interests in decision-making. The only way through this challenge conceptually is to give decision-makers broad discretion to exercise judgment, they say (this blog argues that this needs to be done using the lens of fairness). For a UK lawyer, the model of enlightened shareholder value is a helpful frame for exercising that discretion, however Bebchuk and Tallarita tend to see that as not very different from shareholder primacy. I am not sure many UK boards, especially now they are obliged to report against their directors’ duties to consider stakeholder interests under s172 of the Companies Act, would agree with that assessment.

While the question of enlightened shareholder value is open to ongoing debate, the Bebchuk and Tallarita conclusion does not seem to be. There are fine words in the Business Roundtable’s statement and grandiose expectations for its meaning, but those who signed it do not see it as changing anything:

“corporate leaders don’t contemplate a significant change in corporate strategy…Notwithstanding statements to the contrary, corporate leaders are generally still focused on shareholder value”


‘Stakeholder’ capitalism seems mostly for show, Lucian Bebchuk, Roberto Tallarita, Wall Street Journal, August 6 2020

The Illusory Promise of Stakeholder Governance, Lucian Bebchuk, Roberto Tallarita, forthcoming Cornell Law Review, December 2020

Statement on the Purpose of a Corporation, Business Roundtable, August 2019

What’s the purpose of purpose? Will the focus on corporate purpose deliver real change in the way companies behave?

For all the active discussions about purpose in the context of business, and all the force with which the concept is advocated by its champions, there is a lack of clarity at the heart of the debate. Indeed, too many seem to be seeking deliberately to obscure the fact that there are at least two separate discussions going on. If purpose is actually to deliver on its promise — in the terms used by this blog, fairer business and society — clarity is needed.

Clarity is lacking, though. What’s worse, there seems real disagreement over what a corporate purpose is, and there is even disagreement about what the debate on corporate purpose is all about.

At a recent London School of Economics Systemic Risk Centre event on Sustainability and Systemic Risk, I asked law professor David Kershaw what the distinction is between the new agenda of purpose and the former company law model of articles of association including objects clauses which in effect bar companies from acting ultra vires, beyond their stated scope. This objects clause approach was swept away by the Companies Act 2006 (though it had been rendered largely obsolete prior to this by such clauses being drafted incredibly broadly, giving boards in effect almost unlimited freedom). David highlighted that objects clauses set out what the company was there to do, and purpose statements are about how the company does it.

This idea of purpose as the how is to me an attractive one. But there is active disagreement over which of Rudyard Kipling’s six honest serving men purpose might represent.

For example, Douglas Lamont, the CEO of Innocent Drinks, speaking at the British Academy’s Future of the Corporation Purpose Summit offered his own seductively simple outline of how those serving men operate in framing corporate activity:

“Purpose is the why
Vision is the what
Values is the how”

And Patrick Dunne, former 3i executive, in his magisterial but still highly practical new tome Boards (produced by the excellent Governance Publishing) asserts a further version of this (ironically perhaps, discussed in the ‘Process’ segment of a book he divides into 3 sections, ‘Purpose’, ‘People’ and ‘Process’):

“Vision: What we want to see happen
Purpose: Our contribution to making that vision a reality
Mission: A description of what we actually do
Strategy: How are we going to make it happen”

“Deciding what they [these four plus also brand and culture] should be is one of the most fundamental things for a board and management team to determine,” Dunne says.

The confusion is still worse in some cases, where purpose appears to fulfil multiple roles at once. Take one of the most quoted statements on purpose, Larry Fink of Blackrock’s 2019 letter to investee company CEOs. This shifts from one perspective on purpose to another from sentence to sentence: “Purpose is not a mere tagline or marketing campaign; it is a company’s fundamental reason for being — what it does every day to create value for its stakeholders. Purpose is not the sole pursuit of profits but the animating force for achieving them.” Here, purpose is simultaneously both ‘what’ and ‘why’ it seems. 

But the disagreement goes deeper. It is not just about what purpose is and what it means, but about at what level corporate purpose is being discussed. One of the challenges with business purpose is that there are two very separate conversations going on, and rarely is a clear distinction drawn between them. Indeed, some of the leading proponents of the question of purpose elide the distinction almost entirely. One conversation is about a business’s purpose — what is the driving force for an individual company’s operations — and the other about the purpose of business — what is the overarching role of business corporations in economic life and society as a whole. For example, while the bulk of the focus on purpose in the British Academy’s work on the Future of the Corporation is about individual companies, it does not shy away from brassily huge assertions such as “the purpose of business is to solve the problems of people and planet profitably, and not profit from causing problems”. 

This wide-reaching version of the corporate purpose debate is well illustrated by an article and blog from a pair of law professors from Rutgers, Can a Broader Corporate Purpose Redress Inequality? The Stakeholder Approach Chimera. This sees the debate about purpose purely in the sense of the overall purpose of corporations generally, and seems to take for granted that currently US companies can only consider narrow shareholder interest (a view disputed in Accountable Capitalism). And it views a switch from this shareholder-centricity to a stakeholder model as risky because it is likely further to entrench corporate management: “a stakeholder approach is unlikely to achieve meaningful redistribution of power and resources to weaker constituents and would likely work in the opposite direction. We suggest that a stakeholder approach gives corporate executives both a sword and a shield with which to preserve their advantageous status quo.”

The authors take their cynicism about the Business Roundtable purported restatement of corporate purpose beyond that of the major US investment institutions represented by the Council of Institutional Investors (see The limited responsibility company), and argue that this is not just entrenchment of current power structures but an active distraction from addressing fundamental problems in society: “in our view the entire debate over corporate purpose has so far revolved around the wrong question: it has centered on whether directors are disproportionately focused on shareholders’ interests, neglecting the more important questions of why are weaker constituencies faring so bad in modern-day American capitalism and what can actually be done to elevate their interests”. They believe that these issues would be better solved with more direct interventions, such as changes to labour, tax and competition laws.

This is certainly a worthy debate, but I think focusing the discussion on corporate purpose with regard to the individual company and what it is aiming to do is both more practical and more immediately needed. It should help individual companies prosper and succeed. That certainly seems to be the intent of the FRC in bringing purpose very clearly into the board’s remit in the Corporate Governance Code. Principle B states “The board should establish the company’s purpose, values and strategy, and satisfy itself that these and its culture are aligned.” It does not explain much of what it means by this, other than framing it in terms of long-term success and the business in the context of its stakeholders, and the need to ensure that “policy, practices or behaviour throughout the business are aligned with the company’s purpose, values and strategy”, and that executive pay is also so aligned.

According to David Kershaw and his LSE colleague Edmund Schuster, this introduction of purpose in the Corporate Governance Code is a potential watershed moment, a “purposive transformation of company law”. They suggest this even while noting the FRC’s omission of a definition of purpose, and so the scope for misunderstanding. Again, their focus is on the broader horizon of the overall purpose of the corporation, and they argue that the shareholder-friendly nature of UK company law will frustrate the intent of purpose. They appear to have little trust in the concept of enlightened shareholder value as embodied in s172 of the Companies Act, and the view that a focus on stakeholder interests facilitates long-term value creation for all, including shareholders. Instead they argue that company boards and management will only feel empowered fully to reflect their stated purposes if they have a ‘zone of insulation’ from the usual pressures of the financial markets. In effect, they disbelieve the fine words of institutional investors and think that in practice they will simply favour their own interests as shareholders.

This, though, is surely a chicken and egg situation: shareholders supportive of purpose are drawn to companies with clear, publicly stated and lived purposes, just as much as companies with supportive shareholders are able to develop and assert a purposeful approach. I have argued elsewhere that one reason Unilever was able to survive the bid approach from aggressively managed Kraft-Heinz was because it had previously set out its stall as a long-term business. When Paul Polman started as CEO one of his first acts was to abandon quarterly reporting, encouraging a longer term mindset among shareholders. It put off some investors but encouraged others, meaning the shareholder base was longer term when the hostile approach came. Given the subsequent near-implosion of Kraft Heinz as its zero budgeting model reached its natural come-uppance (as all under-investment will in time), those longer-term shareholders have served themselves — and their clients — well. They also acted in a way that served the interests of Unilever’s employees, customers and suppliers.

Despite the professors’ mistrust, it would seem that institutional shareholders genuinely are ready to embrace purpose, if they believe that managements are themselves committed to them. My friends at SquareWell recently surveyed institutions on the topic, receiving responses from firms with a total of $22 trillion in assets. Not only are these institutions supportive of purpose, more than three-quarters are expecting companies to have set out their purpose:

Squarewell purpose Screenshot
SquareWell: Making Corporate Purpose Tangible

Furthermore, fully 86% expect some sort of declaration that the company has successfully fulfilled its purpose over a relevant reporting period. And they expect purpose to be concrete enough to be measurable: some 75% of these investors believe that KPIs should be set and disclosed with regard to the delivery of purpose, and 59% believe that those KPIs should affect executive pay. Personally I think that this risks attempting to measure the unmeasurable, and certainly it is a further example of the investment community expecting companies to do things they themselves are not yet doing, given that only 22% of the sample have actually themselves incorporated purpose issues into their evaluations of ESG.

Yet in practice uptake of purpose in a formal way has been limited. The SquareWell team report: “In the French market where companies have the opportunity to amend their bylaws to define their ‘raison d’être’, which requires shareholder approval, less than ten companies have taken the leap. Most companies are reluctant as they fear creating new legal risks while they are not convinced of the benefits. The ones which have done so received very high level of shareholder support though.” My informal discussions with investors also indicate a nervousness about the lawyering that is likely to be involved in any incorporation of purpose into bylaws or articles of association, which might risk making purpose meaningless; however, 41% of SquareWell’s surveyed investors believe that the purpose should be so enshrined (though this is fewer than those believing that it is more appropriate for purpose to be asserted through a regular statement in the annual report or otherwise by the board — 55% and 45% respectively).

One example of a French company that has adopted purpose formally was Danone, which at its AGM in June 2020 won no less than 99.4% approval for such a move. Its declared purpose does not seem to have been narrowly lawyered to the extent it is meaningless; indeed, it is general and broad: “The purpose of the Company is to bring health through food to as many people as possible.” Under the société à mission model, the corporate form that Danone has switched to, the company is held to account for delivering on this purpose through the oversight of a special committee, independent from the board (though the rules allow board directors also to be members). This committee’s role is to monitor delivery against the purpose, and to produce an annual report to that effect.

Elsewhere, much has been made by a few commentators of recent IPOs by some B Corps, a corporate model that puts stakeholder interests before those of shareholders. This, they suggest, is a sign of purpose potentially driving changes to business behaviour. Notable amongst these has been insurance business Lemonade, which listed on NYSE at the start of July. “As a public benefit corporation, our focus on a specific public benefit purpose and producing a positive effect for society may negatively impact our financial performance,” states the IPO prospectus in its lengthy Risk Factors section, apparently highlighting just such change. However, sharp insurance industry commentators Oxbow have highlighted just how little of Lemonade’s money actually ends up heading towards its purported purpose: “anything left over is donated to charity in the annual ‘giveback’. This giveback has been a major emphasis in the company’s marketing, but in reality, it’s a very small proportion of actual spending.” So perhaps the legal drafting of risks is just another element of the breezy sales style of the rest of the prospectus, which at times shades into pomposity. Certainly, some investors believe shareholders will make real money from the company, as the share price leapt nearly threefold on debut and even after a drift downwards values the company at some $4 billion. The prospectus reports total 2019 revenues of just $67 million, and a loss of more than $100 million; ‘giveback’ was around $600,000.

A change in business behaviour is the point. It is the purpose of purpose. As the good people at Blueprint for Better Business say in their excellent recent paper Purpose for PLCs: Time for Boards to Focus: “The business needs to be led by that purpose, seeing itself as a social organisation which cares about people and generates the profit necessary to sustain the pursuit of that purpose.” In the Blueprint analysis, purpose is a discipline, a reason to say no to doing things that, while they may be profitable, do not fit with the purpose — the company’s role in society. “The shift is from a narrow focus on financial value creation alone to a wider, richer view of what a company exists to do in the world.” To my mind this paper is one of the clearest contributions to the debate, short and usefully to the point. It goes on:

“While some stakeholder relationships are more material to business success, being purpose-led means seeking to have a positive impact on all those whose lives are touched by the company, avoiding manifest unfairness to anyone.”

This is the purpose of purpose, unlocking the opportunity of positive impact through business — and perhaps more importantly, having a reason to avoid negative impacts, even if they are profitable in the short term. For all of the cynicism that the noise around corporate purpose evokes, not least because so many users of the term seem to have no real intention to change anything that they do, and all the disagreement about what it actually means, there is a big opportunity for positive impact. Establishing a corporate purpose does need to change what a company does, perhaps on all the dimensions of what, why and how. By each company seeking to have a purpose that goes beyond the simple generation of short-term profit but that instead looks to generate long-term prosperity, we are more likely to achieve that economic prosperity. A broader and larger economic prosperity can then be more fairly shared between stakeholders. Perhaps we should do that on purpose.

See also: What’s the purpose of purpose? (II)


Purpose for PLCs: Time for Boards to Focus, Blueprint for Better Business, 2020

The Purposive Transformation of Company LawDavid Kershaw, Edmund Schuster, LSE Law, Society and Economy Working Papers 4/2019

Boards, Patrick Dunne, Governance Publishing 2019

Profit & Purpose, letter to CEOs 2019, Larry Fink, BlackRock

Principles for Purposeful Business, How to deliver the framework for the Future of the Corporation, The British Academy, 2019

Can a Broader Corporate Purpose Redress Inequality? The Stakeholder Approach Chimera, Matteo Gatti, Chrystin Ondersma, forthcoming Journal of Corporation Law, 2020
Also Oxford Business Law blog, June 11 2020

Making Corporate Purpose Tangible, SquareWell, 2020
Also Harvard Law School Forum on Corporate Governance blog, June 19 2020

Who’s drinking all the Lemonade?, Oxbow Partners, July 30 2020

Lemonade IPO Prospectus, 2020

Poverty isn’t bashful

Poverty is bashful says the Pope repeatedly in the first of the BBC’s Rethink essays. Rethink is a series of podcast reflections on the recovery from Coronavirus, and the world that we might aspire to creating once we escape our current constrained existence. Many hope that this is an opportunity to build a fairer world.

Far be it from me to suggest that, contrary to Catholic doctrine, the Pope is fallible — and I must admit that this particular Pope has far more direct experience of poverty than I do. So I won’t say that he is wrong; I will say that I disagree. Poverty is not bashful.

Poverty isn’t bashful. Rather, it is forced to the sidelines of our lives, it is ignored, it is chased from our streets by officious authority, squeezed into slums and ghettos, moved from public space by private closures — the gated communities of the wealthy or their ungated analogues POPS (the informal acronym for the privately owned public spaces that increasingly crowd our cities). It is hidden because it is inconvenient and difficult and so rarely reaches public or media attention. Companies hide it away by casualising their own workers, or outsourcing services or manufacturing, in effect pushing poverty down their supply chains, while squeezing what they pay suppliers making poverty wages more and more likely. That makes poverty less visible but no less present.

Poverty is forced into the corners because we hide it away, preferring not to know, preferring not to be embarrassed by the shame it inflicts on our comfort in the face of its discomfort. We collectively turn our eyes away, just as we do from the beggars on our streets or the sellers of the Big Issue (and other street papers around the world). In our comfortable lives it is more reassuring not to think of those on the margins, those who struggle day-to-day. These are not just the homeless, but also those struggling to put food on the table, dependent on free school meals and foodbanks — people whose ranks were growing ahead of the Covid crisis and have swollen still further during it. 

Poverty is not bashful. We choose to ignore it.

In part, we ignore poverty because we tend to fool ourselves that somehow most poverty is deserved. While some — perhaps — is, much arises from unavoidable circumstance. We so love to believe in meritocracy that we tend to ignore this unavoidable nature of much of the poverty around us.

One benefit of Covid 19 is that it has become harder to ignore others, it has become harder not to see the poor. Instead, it has become more clear how connected we are, how intimately our very chance to live is tied up with others. In a similar way, it has become harder for businesses to ignore the interests and basic needs of their workforce and the implications of how they are paid and treated. It has become almost a cliche among stewardship-minded investors that they will need to focus more in the next several months on the S (social) in ESG — meaning that they will increasingly challenge the companies in which they invest to do better in terms of how they treat their workforces. In part I suspect the reason for this new focus is guilt that they have previously given these issues only limited attention.

The Pope says we need to see the poor, that currently we don’t see them because we perceive them as only part of the landscape, just things. This I do agree with: we need instead to recognise their humanity. Companies need to recognise the human needs of their employees, and of the workers in their supply chains also. Just seeing more clearly would deliver to us all a greater sense of community, and would deliver a real opportunity to reduce the poverty and unfairness around us.

That would be a great, and a fairer, legacy of the Covid crisis. Let’s not be bashful.