I have lived experience of safetyism, and it is dangerous. By the ages of 88 and 90, my parents had lived in the same house for 48 years. Frail but mentally unimpaired, they wanted above anything to remain there, so I had arranged periodic visits from carers sent by the local council. One morning two social workers turned up—evidently a check-up. Finding them in bed, and seeing their fragile condition, they realised that an emergency might happen at any time. Having visited them shortly before, I can only agree.
But it was a Friday: what if an emergency happened that weekend when no visits were scheduled? And who knew best, my parents or these two professionals? Would the council, or these social workers themselves, be liable? With bold overconfidence, they assumed agency, infantilising my parents by ordering them to pack, bundling them into their car, and depositing them at a care home.
Why did they do it? A tick-box rule designed to eliminate risk to the department, and hence liability. As next of kin, they phoned me—I was in America giving a lecture. They reassured me that nothing was wrong, it was purely precautionary, the care home was wonderful—“like a luxury hotel.” But my parents had never been in a luxury hotel: they valued privacy and their simple lifestyle. I flew back, reaching them the following afternoon.
“Why are we here?” my mother’s voice still echoes. “We want to go home.” I should have driven them home that instant, but others who have done that have been prosecuted; in any case we were instinctively compliant with authority. Instead, on the Monday morning, I phoned the social workers and came to an understanding that they would take my parents home that very day.
Bureaucracies can work fast when avoiding liability, as this one did; they are less energetic at correcting mistakes. On the Tuesday, my mother awoke still in the care home. She fell out of the unfamiliar high bed, and broke her hip. On the Wednesday she died, with me beside her. My father returned home, and lived for another seven years. My mother had survived in her home for 48 years; safetyism took only five days to kill her.
What was going on here? It is the inability of a bureaucracy to cope with the inevitable uncertainties that abound in our lives. These are open-ended situations in which we cannot know enough to be certain what the best thing is to do, which—in the new science of “radical uncertainty”—are known as “large world” problems.
Those social workers had reduced their situation to a “small world” problem: a more bounded situation in which it was hoped they might achieve, at least for themselves, complete certainty. The trouble is that this approach amounts to wishing unknown unknowns away. Closing your eyes to that which you can’t see coming cannot eliminate the dangers, it can only ever shift them elsewhere. The social workers may have reduced the council’s liability for under-reacting to one sort of risk, but—clearly—they catastrophically increased another one for my parents.
In spring 2020, an analogous reduction of a large world problem to a small one cleared hospital beds for Covid-19 by moving 25,000 patients into care homes. In freeing up NHS beds, one risk was eliminated: hospitals that couldn’t cope. But disastrously, the overall risks of mortality were shifted to care homes, which became killing fields.
From cradle to grave
The defining feature of safetyism is the displacement of judgment by procedure. This goes hand-in-hand with the treating of large world problems as small world problems, because the known, specific and quantifiable risks of the latter can be put on a tick-list and approached procedurally in a way that unbounded uncertainty never can.
It is quite a distinct phenomenon from the supposed “health and safety culture” that libertarians bemoan
for curtailing the right of individuals to run risks. The question there is the balance between freedom and danger: though contested, such paternalism really is about trying to keep people safe. Safetyism, by contrast, does nothing to protect the citizen. The only thing it keeps safe is the bureaucracy. And, sadly, my own experience with it does not end with my parents.
Another of my brushes with it began when a social work department removed the toddler of a young married couple due to reports of neglect. The home was somewhat chaotic, but not an extreme case, and there was no evidence of abuse. Once in foster care, the girl refused to eat and lost weight drastically. Nevertheless, when the couple are expecting their second child, the social workers decide that the baby should be removed at birth: they tell the mother at seven-months pregnant. Ten minutes after birth, the new baby boy is removed from her arms and placed with other foster carers, separate from his sister. After 19 months, still in “care,” he can neither walk nor talk.
At that point, and with the two children not even knowing they are siblings, as distant relatives my wife and I use a new legal loophole to bypass adoption, which had been made phenomenally difficult by safetyism (only a very low proportion of children in care make it to adoption each year). We become their special guardians. In being assessed for suitability, the entire focus is on their physical safety. “Do you remove all your plugs every night?” Our mystified response—“Do you?”—is the wrong answer. We agree to replace windows with shatterproof glass, and our home becomes festooned with fire extinguishers.
The children are thereby deemed “safe,” but step back and ask what has been achieved. The first child, anxious to the point of malnutrition, had been placed at risk of stunted development (or worse—after she had been released from “care,” it transpired that seven other young children had recently died while being fostered for this local authority). The second child, while a foetus, had been in the womb of a traumatised woman dreading the removal of her baby son from her arms. Again, we are on solid ground: research shows that stress in late pregnancy inflicts irreversible damage to the foetus.
My experience is far from unique—here is another story that was closely reported in the serious press, which reveals a similar logic. A policewoman turns her back, hears her baby fall over and cry, and takes him to hospital where a broken leg is missed. Still worried, she brought him in again where the diagnosis was finally made, but a child protection inquiry was also triggered. It led to X-rays which were misread, with ordinary signs of infant bone growth somehow mistaken for additional fractures. The professionals involved may have honestly believed that, but they couldn’t honestly claim to know what caused them: this baby is unfortunately a large world problem. But safetyism has converted him into a small world, tick-box problem—and the box marked “definitely not” cannot be ticked. Tick-box removal of the baby from her parents follows.
That mother is relatively fortunate: she gets her baby back 10 days later, albeit parents and baby all traumatised. But in a similar case, also well reported, the mother is not so fortunate. A bruise on the child is reported to a clinic by the mother. Following her local council’s exceptionally “safety-alert” rules, the health worker refers the baby to the local hospital. There, another “safety-alert” doctor misdiagnoses an X-ray, triggering the removal of the baby. The medical error is corrected by a specialist, and belatedly the hospital returns the baby to the parents. But the social worker refuses to accept the medical judgment, requiring the police to remove the child into “care.” It takes 16 weeks and three newspaper-financed legal battles before a judge orders her return to the now-shattered parents. We are at last on solid ground: research shows that separation damages attachment at a critical stage, with lasting consequences. Again, safetyism privileged the elimination of risk to the social work department, fussing over a vanishingly small risk to the baby for which it could be held liable, at the cost of an overwhelming likelihood of inflicting material harm. Only human judgment saved the day.
Had either situation been viewed as a large world problem, the social work department might instead have tried talking to and, if needs be, supporting the parents involved, as is standard in the Netherlands. There are now 80,000 British children in “care,” a number inflated by safetyism: see Hilary Cottam’s Radical Help (2018) for a devastating critique of the system.
Stumbling into safetyism
So how did we get here? Safety is second in Abraham Maslow’s hierarchy of needs: only sustenance comes before it. Like other mammals, we crave it, but we are atypical, endowed with creative imagination that has taken us from the savannah to modernity. The price of repeatedly imagining a better life is that we sometimes strive for things whose consequences we cannot foresee or understand.
And herein lies the rub: within our complex, interdependent, radically uncertain world, absolute safety is unattainable. If we insist that some outcome be made certain, the irreducible uncertainty of the system will surface in increased exposure to risk in some related outcome, perhaps one where it does more damage.
Because we live up to our necks in uncertainties, we have evolved wisdom to check the lure of the dream of total safety. Rather than cower in shelter, we work and send our kids to school; we take holidays; we shop. We understand, because we have to, that if we all hunkered down in permanent lockdown, we would be safe until we starved.
Uncertainties are perennial to our lives, as they always have been. What is new is that many of the decisions that affect them have been captured by bureaucracies, who do not and maybe cannot develop similar practical judgment, as another story elucidates.
An elderly man lies in hospital following a stroke. His life’s work has been to build a modest business, but he has neglected to make his wife a co-signatory of his account. Bills need to be paid, contracts signed. His cognition has been affected, and he cannot concentrate sufficiently to run the business from his bed. He realises his wife must take over and asks her to bring a lawyer to the hospital for the necessary power of attorney.
A lawyer comes. She is good, experienced and meticulous, as well as an acquaintance of mine: this is her story. While in hospital a grant of attorney needs medical approval. The hospital psychologist worries that a decision about approval might open the hospital up to liability in the event of the power being misused. She tells the lawyer to seek approval from the Court of Protection, which can transfer the power of signature to the wife regardless of the man’s state of mind. She thereby protects the hospital.
But the Court has a year’s backlog of cases, so the business will collapse. In their distress, the man and his family see that their only option is to take him home prematurely, where the lawyer certifies that the man can sign over the power. The hospital is safe, but in protecting itself it has increased the risk to the man’s life.
It is not only individual bureaucracies but whole professions that behave like this. They dig a moat against outsiders. Professions are defended by inside knowledge as they set their own standards, and use them to exclude outsiders with more vigour than to enforce their rules on their members.
Decisions in business and government are shaped by actuaries, lawyers, accountants and regulators. And the management of pension savings, illness and family stresses in our own lives are increasingly taken by fund managers, medics and social workers. Their specialist techniques can sometimes improve decision-taking, but not if they are deployed as a substitute for all judgment. Yet that is what is happening: bureaucracies and siloed professional bodies sidestep the practical trade-offs that the journey through quotidian life forces us all to make. Their narrow formulaic decision-taking protects their members from embarrassment and liability by shifting all the myriad uncertainties that can’t be seen coming on to other people.
Our best shot at managing such uncertainty is through the entity we now recognise as the most complex in the known universe: the human brain. Even in chess, a game whose problems are defined by a short set of unambiguous rules and limited choices, human judgment is important. Although AI now reliably beats human players, real players aided by AI turn out to beat all pure AI contestants.
And real world decisions are incomparably more complex than any game: rules are ill-defined, options for actions are unbounded—and so too are dangers. When human judgment is displaced in navigating our stubbornly surprising reality, things inexorably go wrong. If you still doubt it, look at the ruin that safetyism has very likely wreaked on your retirement.
Defined dangers
Millions of us used to enjoy defined benefit (DB) pensions, with reliable payments typically calculated as a proportion of final salary. They provided security around which life could be planned. Nowadays, they have largely disappeared. They went because several professions each found small world solutions within one large world problem. What might a DB pension mean? It should mean that a company—and the government—will do all they can, subject to balancing other priorities such as the survival of the business. Such caveats should not be controversial: even the “fundamental liberties” in the European Convention on Human Rights are expressly allowed to be balanced against the wider public good.
But corporate lawyers advised boards to interpret defined benefits as legal guarantees: pensioners should have absolute certainty. This shifted all liability to firms. Actuaries then did alarming demographic sums about worst-case liabilities. Asset managers addressed the resulting problem: what portfolios could be certain to meet these liabilities in worst-case scenarios, such as what would happen to my pension if all companies and universities in Britain closed tomorrow? (To which the true answer is nothing, because a meteor must have destroyed the Earth.) Finally, the regulators addressed a fourth: what assets were sufficiently safe to be certain to maintain their value, day-by-day?
“Lawyers defined pensions as guarantees; actuaries did alarming sums; asset managers fixed on worst-cases. So pensions got too costly—and closed”
The cumulative consequence of this sequence is that a pension fund must hold ultra-“safe” assets that yield virtually no return, so that huge contributions are needed to guarantee the pension. Defined benefits became so absurdly expensive that companies closed them. In consequence, your only option might well be a “money purchase” pension, whose value will depend on unpredictable annuity rates and the unknowable closing value of your fund of contributions. Thanks to safetyism, your pension will be smaller, and riskier, than necessary.
Meanwhile, our successive pensions ministers, who should have co-ordinated a collaborative sharing of uncertainties, watched in bewilderment, perhaps ill-advised by a “small world” civil service. In normal times, a well-invested company pension should generate enough annual revenue to match annual outgoings—notional and volatile “snapshot” valuations of assets and liabilities merely confuse the matter. In abnormal times such as during a global crash, the government can and should stand behind companies, using its control of future taxes as the shock absorber.
Lucky me: I am one of the few with a guaranteed DB pension, underwritten by your taxes. In other aspects of life, sadly, I remain just as exposed to the dangers of safetyism as anyone else.
Pushing back
By 2019, the mother of the two children I had become a special guardian for was pregnant with her fourth child (after having been banned from seeing the third). She was living with a new partner and desperately wanted to keep a child.
We were phoned by social services to ask whether, once born and removed, we would again be special guardians: we managed to stay polite. Mercifully, safetytism tied itself in Covid knots. Since social workers must be protected, they could not visit to inflict “safety” on the baby, so she stayed with her parents. Now a toddler, she is doing fine, and her mother has made real efforts to turn round her life.
Safetyism downplays the possibility of such positive outcomes, emphasising specifically the foreseeable worst case to the exclusion of all others. It is sometimes well-intentioned but never takes a rounded view. Individual potential dangers make the tick list, and are afforded overwhelming priority over the bigger picture. So safetyism is not innocently foolish: it is the repeated retreat from judgment that is cumulatively dangerous.
“Safetyism downplays positive possibilities, emphasising specifically the foreseeable worst case to the exclusion of all others”
It is rationalised by a seductive mantra of “if one life is saved,” which in practice means “if one life is saved through a particular channel of danger we have anticipated, and for which someone could conceivably hold us responsible.” And this comes at the expense of the (often life-saving) advantages of making sure that dynamic situations are managed with a comprehensive and flexible sense of responsibility.
Safetyism is still rolling across the landscape, but it can be rolled back. The mortal shock of a pandemic—a novel virus that arrived out of nowhere, and that nobody initially knew how to deal with—has propelled uncertainty decisively into lives. I illustrate the effect on the culture with quotes from a newspaper that sits on my desk as I write. From a letter—“Keats used the term ‘negative capability’ to describe a human acceptance of uncertainty”; from an article by a monk on what meditation teaches—“enduring radical uncertainty”; from a review of a jazz masterpiece—“teaches us about accepting uncertainty.” Just as Covid taught us to Zoom, it has taught us to be wary of
know-alls.
Turning to the world of letters, in The Upswing (2020), American sociologist Robert Putnam shows that the change happens when powerful new ideas coincide with a mass shock that illustrates them. If Covid is the shock, the ideas are now there too. Michael Sandel is developing a notion of “contributive justice,” according to which it is unethical to strip people of agency: we must all take moral responsibility. John Kay and Mervyn King’s Radical Uncertainty argues that human judgment is essential because large world problems abound. These ideas are demolishing the basis for the whole architecture of delegated decision-making by siloed specialists.
In a democratic society, when we understand that something has gone amiss, we can all set about changing it, decision by decision. And if you work in one of those siloed professions yourself, you can use its association to challenge current practices. Actuaries have already started, with good reason. In February 2020, actuaries reported to one of our largest life assurance companies on its risk profile: an ocean of demographic statistics that were readily quantified, but no mention of pandemics.
Only one year later, those statistics were make-believe: a large world phenomenon had changed deaths and births beyond the range of its small world modelling. To its credit, the Institute and Faculty of Actuaries took action—inviting John Kay, Andy Haldane and me to address their membership. We each gave a lecture, followed by a joint discussion. You can find it online.
Our message was not comfortable. When the “feedback forms” were compiled, nobody in an audience of several hundred rated my performance just “satisfactory.” The reactions were all clustered at the two extremes. Encouragingly, acceptance outvoted denial by five-to-one. Actuaries are building a new consensus for change; equivalent honesty needs to happen profession by profession.
You can keep your head below the parapet—another little retreat into safetyism. Or you can speak out: some will disagree, as we found. But only by speaking out will we get change. The price of the good life, including protection from the tyranny of self-serving bureaucracies, is not merely “eternal vigilance.” It is courage.
