Every manager is a health and safety issue
In brief: Most organisations say they care about staff wellbeing, but the real test is how work is actually designed and managed. Drawing on research from humanitarian workers, emergency responders and mainstream workplaces — as well as my own experience managing teams — this piece explores how leaders can become either a source of psychosocial harm or a protective factor, why chronic overload should not be confused with high performance, and what it means to build workplaces where people can work hard without being slowly consumed by the work.
Many years ago, while I was working in Iraq, I stumbled across a survey of international aid workers asking them what caused them the most stress.
I remember trying to guess the answer before reading it. Perhaps it would be the insecurity: the possibility of bombings, shootings, kidnapping or simply being in the wrong place at the wrong time. Perhaps it would be living thousands of kilometres from partners, children, families and friends, or the constant cultural adjustment, the unfamiliar food and living conditions, and the strange intensity of often living and working in the same small compound with the same people, day after day, whether you liked them or not.
But the answer I remember was much more mundane.
Their managers.
I no longer have the survey and have not been able to relocate it, so I am not going to pretend that I remember its numbers accurately. But I clearly remember the findings, partly because it seemed so improbable. How could your line manager possibly be worse than living and working in a war zone?
As it turns out, the finding was probably much less surprising than I thought. A later study of 218 humanitarian workers living across 63 countries found poor management to be the third most frequently cited source of stress overall. Respondents described frustration with incompetent or unfair supervisors and authoritarian styles of management, and with systems in which people felt neither protected nor heard and could not challenge what was happening.
I was fortunate to encounter that idea relatively early in my career because, as my own responsibilities grew over the years and I found myself managing larger teams and eventually organisations, it remained somewhere in the back of my mind. Whether I intended it or not, whether I was conscious of it or not, I was part of the working environment experienced by the people I managed. I could make that environment better. I could also make it considerably worse. And I think leaders, particularly senior leaders, routinely underestimate just how much power that gives us.
Last week I wrote about why we are so bad at caring for ourselves, and argued that those of us with authority over other people's workloads, wages, schedules and leave carry a responsibility to change the conditions under which others live and work. This article follows that thread into the workplace, where so many of those conditions are made.
The injuries we don't normally associate with leadership
I was reminded of all this recently by a new Australian study of emergency responders.
Researchers Greg Drummond, Joseph Cuthbertson, Kate Baecher and Matthew B. Thompson surveyed 54 frontline emergency workers about moral distress and moral injury. The sample is small and the research exploratory, so we should not turn it into a sweeping claim about every responder in Australia, but its findings are nevertheless striking.
The organisational environment repeatedly emerged as a source of moral distress and injury. Respondents described distrust in leadership, perceived procedural injustice, bullying, leader detachment and, perhaps most interestingly, a discrepancy between the values their organisations claimed to hold and those actually demonstrated through their decisions and behaviour. Empathetic and supportive leadership, meanwhile, emerged as a protective factor.
These are people whose jobs can expose them to burning homes, terrible injuries, death, grief, violence and other human beings on what may be the worst day of their lives. Yet some of the deepest distress they describe comes less from the traumatic event outside the organisation than from what happens inside it.
Moral injury in particular helps explain why. People can cope with extraordinarily difficult circumstances when they understand why they are there, trust the people alongside them, believe they are being treated fairly and feel that the organisation they serve broadly acts consistently with the values it proclaims. What can become particularly corrosive is being asked to tolerate or participate in something that violates those values, or to stay silent about it, while being told that the organisation stands for something quite different.
But this is not really an article about aid workers or emergency responders, because when you begin looking at the broader workplace literature, the same pattern appears. A large meta-analysis of longitudinal research involving more than 30,000 employees found that destructive leadership had subsequent detrimental effects on important employee outcomes including job satisfaction, commitment, avoidance and workplace behaviour. A separate systematic review of empathetic leadership identified effects spanning employee wellbeing, interpersonal relationships, attitudes, equity and performance, though the research base remains uneven.
This reaches well beyond whether people like their boss and into the territory of health.
(Picture by Sydney Latham)
Your manager is part of your working environment
Australia's work health and safety framework increasingly gives us a useful way of thinking about this. We have become much more explicit about psychosocial hazards: aspects of the design or management of work that can cause psychological, and sometimes physical, harm.
Safe Work Australia lists among them excessive job demands, low control, poor support, inadequate recognition, unclear roles, poorly managed organisational change, poor organisational justice, bullying, harassment and workplace conflict.
Look carefully at that list and something becomes obvious: managers influence almost every item on it. They influence how much work people are expected to do, how much control they have over doing it, whether priorities are clear, whether people receive adequate support, how mistakes are handled, how recognition is distributed, how change is communicated and whether decisions are made fairly. Poor organisational justice itself includes inconsistent decision-making, inadequate communication, blaming people for things outside their control and failing to deal appropriately with harmful behaviour.
This points towards a shift in how we might understand the role itself. A manager is more than the person responsible for getting work done. A manager is part of the psychosocial environment in which work gets done, and depending on how they exercise that role, they can become either a protective factor or a source of risk.
Of course, we need to be careful here. It would be easy to turn this into another attack on middle managers when they are very often caught inside precisely the same system. A line manager may be demanding impossible outputs because their regional manager is demanding them. The regional manager may be responding to an executive target. The executive may be responding to a board mandate or investor expectations. Stress, like culture, cascades down a hierarchy, so responsibility must cascade upwards as well.
It is hardly reasonable for a board to impose unrealistic targets and chronic understaffing, punctuated by continuous restructures, and then send the middle managers who have to implement those decisions to a half-day seminar on empathetic leadership. If we are serious about psychosocial safety, we need to examine the system producing the behaviour before we examine the behaviour of the person closest to the employee.
Australian research on what is known as Psychosocial Safety Climate is particularly useful here. The idea, put simply, is that employees learn whether their organisation genuinely prioritises psychological health and safety by observing what managers actually do, particularly when wellbeing competes with other priorities. In a longitudinal study involving 49 team leaders and 281 Australian healthcare workers, leaders who placed a high priority on psychological health helped build stronger psychosocial safety climates, while leaders who did not could erode an initially healthy one.
The operative word is actually.
Almost every large organisation today has values, and many have wellbeing strategies, EAPs, flexible-work policies, mental-health initiatives, resilience programs, employee surveys and statements about putting people first. But employees learn what an organisation really values when something has to give.
What happens when the deadline cannot be met without people working all weekend, when somebody says the workload has become unsustainable, when an employee challenges somebody more powerful than they are, or during a restructure? What happens when hitting the quarterly target conflicts with protecting people's wellbeing? That is when organisational values stop being words and become evidence.
The cult of being “hardcore”
There is, of course, a radically different school of management, and perhaps its most visible contemporary exponent is Elon Musk.
When Musk took control of Twitter in 2022, remaining employees received an ultimatum. Building the new company, they were told, would require an “extremely hardcore” culture involving “long hours at high intensity”, with only exceptional performance considered sufficient. Employees who did not want to sign up to that environment could leave.
There is a weird seductive logic to this. Great things are difficult, and building rockets or global technology companies requires extraordinary effort. People pursuing something genuinely ambitious will sometimes work very hard. I know this from my own life and, if I am truthful, I have sometimes expected too much of myself and probably of others because I cared deeply about what we were trying to achieve.
The problem begins when exceptional effort stops being exceptional and becomes the operating model. There is a significant difference between voluntarily throwing everything you have at something for a short period because the moment demands it, and designing an organisation around the assumption that human beings should routinely operate at maximum intensity. It is the difference between sprinting when necessary and running a marathon in which everyone is instructed to sprint.
Musk has also attracted attention for a more recent argument about empathy, describing what he sees as an “empathy exploit” and at one point calling empathy a fundamental weakness of Western civilisation. Fairness requires acknowledging the rest of his argument: he has also explicitly said that he believes empathy is good and that people should care about others, while arguing that empathy can be manipulated or applied without sufficient attention to broader consequences.
There is actually a legitimate philosophical question buried in there. Empathy alone cannot determine good policy or good management. We can empathise deeply with one person while failing to understand the consequences of a decision for hundreds of others. Leaders sometimes have to say no. Sometimes they have to close programs, restructure organisations, dismiss people or make decisions that cause genuine pain.
But none of that requires abandoning empathy. Empathy is not the same thing as agreement, and it does not require indulging people or avoiding difficult decisions. At its most useful, empathy is information: the capacity to understand how a situation looks and feels from somewhere other than where we are standing and acting accordingly. That makes it a form of organisational intelligence, and one that counts for most when we hold power.
What power does to perspective
I have no interest in diagnosing Elon Musk, Jeff Bezos or any other executive from afar. We do far too much amateur psychology of public figures already, and I do not know these people. The more interesting question is what certain approaches to management reveal about the assumptions we can develop when we hold considerable power.
One assumption is that because I am prepared to sacrifice almost everything for the mission, asking everyone else to do so is reasonable. Another is that people who cannot tolerate the environment simply aren't sufficiently committed or talented. Another is that measurable output tells us more than subjective human experience. And perhaps the most consequential is that human beings can gradually become ‘inputs’ into the achievement of an ‘outcome’ rather than people whose lives exist far beyond the organisation employing them.
Power itself may make this easier. Classic experimental research found that people induced to experience greater power were less likely to spontaneously take another person's perspective and were less accurate at identifying others' emotional expressions. Like much psychological research, these experiments have their limits, and they point towards a plausible problem all the same: power can increase social distance, and social distance can make the consequences of our decisions less visible to us.
The antidote is a different practice of leadership rather than less of it. Research by Tsedal Neeley and Sebastian Reiche examining 115 senior global leaders identified what they called downward deference: powerful leaders deliberately reducing social distance and deferring to people below them when those people possessed expertise or relationships the leader did not have.
I love that concept because it captures something easily lost as we rise through organisations. The more senior we become, the more information reaches us already processed. Somebody summarises it, somebody cleans the spreadsheet, somebody turns the complexity into three PowerPoint bullets, and somebody colours the dashboard green. Eventually, we can become immensely powerful while knowing surprisingly little about what it actually feels like to work inside the organisation we run. Good leadership therefore requires deliberately closing that distance.
When pressure becomes a system
Amazon provides a different and perhaps even more interesting case because the issue is less about the personality of one famous chief executive and more about the systems through which work itself can be engineered.
In 2024, a US Senate committee investigation alleged that Amazon's intense focus on warehouse productivity and speed was contributing to workplace injuries, and that the company had failed to implement some internal recommendations that might have reduced injuries because of concerns about productivity.
Amazon strongly disputes that characterisation. It argues that its productivity expectations are reasonable and that some of the research relied upon by the committee was flawed, and it points to substantial improvement in its injury rates. The company's most recent figures report a 43 per cent improvement in its global recordable incident rate since 2019 and a 70 per cent improvement in its lost-time incident rate, alongside more than US$2.5 billion invested in workplace safety. Those are company-reported figures, but they are significant and should be acknowledged.
My point is therefore not that Amazon is uniquely evil, or even that we can settle every disputed claim about its warehouses from afar. The case illustrates a much broader question facing modern organisations.
What happens when almost everything that can be measured is optimised, from time and movements to units produced, calls answered, patients processed, reports completed, utilisation rates and billable hours, while the human experience of performing the work remains considerably harder to quantify? What gets measured tends to acquire power, and what cannot easily be measured can slowly disappear from decision-making.
At its worst, organisational wellbeing then becomes strangely inverted. We create working conditions that exhaust people and subsequently teach them resilience so that they can tolerate those conditions better. We offer mindfulness to people whose workload is impossible. We provide an EAP to employees whose manager controls every minute of their day. We run a mental-health awareness week while continuing to reward the executives producing the system that is making people unwell.
There is nothing wrong with mindfulness or counselling, and nothing wrong with resilience; I work in resilience myself. But resilience should never become a technology for helping people endure preventable harm.
Compassion and performance are not opposites
Perhaps this is where I depart most strongly from the cult of the hardcore leader: I don't believe that compassion and high performance sit at opposite ends of a spectrum.
I have met very compassionate leaders with exacting standards. I have also met leaders who confused aggression with courage, exhaustion with commitment, fear with respect and obedience with performance.
Compassionate leadership is often misread as permissiveness, as letting people do whatever they want. What it asks of us is harder: to care about the human being while dealing squarely with the problem.
You can tell someone their performance is inadequate and still preserve their dignity. You can reject someone's proposal while taking their concerns seriously. You can restructure an organisation while being frank about what is happening and why. You can expect people to work hard without expecting their work to consume their entire lives. You can make decisions people hate and still make them fairly.
In fact, I would argue that empathy grows in importance as decisions grow harder. The leader who understands the human consequences of a decision may still make exactly the same decision. But they are likely to implement it differently, communicate it differently, mitigate its consequences differently and perhaps recognise costs that would otherwise have remained invisible. Far from relieving leaders of responsibility, empathy makes us better able to carry it.
So what should leaders actually do?
If we accept that leaders form part of the psychosocial environment experienced by their staff, there are some fairly practical consequences.
First, treat leadership as a psychosocial control as well as a performance function. When organisations assess managers, they should examine what their teams delivered alongside the conditions under which those results were produced. Chronic turnover, excessive overtime, unresolved grievances, repeated sick leave, fear of speaking up and striking differences in wellbeing between comparable teams are management information as much as they are HR problems.
Second, fix work before trying to fix workers. If the workload is structurally impossible, no amount of yoga will make it reasonable. If the problem is low control, improve control. If the problem is role ambiguity, clarify roles. If the problem is understaffing, examine staffing. If the problem is injustice, change the decision-making process. Safe Work Australia's guidance is quite explicit that psychosocial risks should be addressed through the design and management of work itself.
Third, stop assuming that technical excellence automatically qualifies somebody to manage people. The best engineer does not necessarily make the best engineering manager. The best salesperson may be a terrible sales director. Laurence Peter warned of this more than half a century ago with what became known as the Peter Principle: in a hierarchy, people tend to be promoted on the strength of their current job until they reach one they cannot do, and so rise to their level of incompetence. Leadership creates an entirely new category of responsibility: responsibility for other people's working environment. Selection, promotion, development and performance assessment should reflect that.
Fourth, train managers, but also give them the authority, the resources, the time and the standing to manage well. Telling a supervisor to support their team while giving them 25 direct reports and impossible targets, without any real discretion, is organisational theatre. Managers themselves need manageable work, clarity, support and psychological safety. Otherwise the pressure simply travels downwards.
Fifth, make organisational justice a leadership discipline. People will not always agree with decisions. They can nevertheless understand how decisions were reached, have opportunities to be heard, know that rules are being applied consistently and expect to be treated with dignity. Fair process matters enormously, particularly when outcomes are painful.
Sixth, measure the distance between stated and enacted values. Ask employees where they see the organisational values contradicted, rather than only whether they can recite them, and ask what happens when living by those values becomes expensive or threatens a target. That is where culture becomes real.
Seventh, deliberately reduce executive distance. Senior leaders need mechanisms through which they encounter the organisation without everything being filtered through management layers. Talk to frontline staff without their manager standing beside them. Listen to people who disagree. Visit the difficult sites as well as the showcase ones. Ask what senior management does not understand about how work actually happens.
And finally, remember the asymmetry of leadership. A passing comment from a CEO can become months of work for hundreds of people, and a poorly considered target can restructure thousands of working days. An executive may forget a conversation five minutes after leaving the room while the person on the other side thinks about it all weekend. Power amplifies us: our strengths travel further, but so do our blind spots.
What organisations do with care
The World Health Organization's guidance on mental health at work points in the same direction as the Australian framework, identifying risks such as excessive workloads, understaffing, inflexible hours and job insecurity, and placing strong emphasis on changing the way work itself is organised. At Plan C, the community resilience charity I chair, we have recently tried to put some of this thinking into practice through our Mental Health and Wellbeing Policy. We are a small organisation, largely staffed with part-time staff, often short of resources and working on issues that can be emotionally demanding, so these pressures are very real for us.
The policy starts from the idea that wellbeing is partly shaped by the way work is designed. Staff have responsibilities too: we expect people to speak up when workloads are becoming difficult, respect agreed boundaries, use available support and raise patterns that cannot continue. Plan C, in turn, carries responsibility for making those behaviours possible.
That quickly becomes practical. Work should be planned around people's contracted hours rather than an assumed willingness to keep doing more. When an important new task arrives, there should be a conversation about what existing work will be delayed, reduced, reassigned or dropped. If someone repeatedly works outside agreed hours, that should lead us to look at workload and resourcing rather than simply allowing it to become normal.
The policy also provides preventive wellbeing days, reproductive health and wellbeing leave, workplace adjustments and access to a confidential Mental Health and Wellbeing Lead. We ask managers to respond early when someone appears overloaded, perhaps by changing priorities, moving a deadline, redistributing work or encouraging leave.
There are larger questions underneath these practical measures. Can someone say they are struggling without worrying that their competence will be questioned, and do part-time staff genuinely have permission to work part-time? Does a manager express concern about an exhausted employee and then leave the workload unchanged, and does an organisation praise the person who repeatedly rescues impossible situations without asking why those situations keep happening?
Plan C does not always get these things right, and writing the policy has not solved them., but it gives us something more concrete against which to examine how we actually behave. I increasingly think workplace wellbeing lives in these ordinary decisions: how work is allocated, what deadlines are set, whether enough people are employed, how meetings are run, what happens outside working hours and what kinds of behaviour leaders reward. Those choices tell us far more about an organisation's approach to care than the words on its wellbeing page.
* * *
I still think about that survey I encountered in Iraq. At the time, it seemed bizarre that people surrounded by the risks and discomforts of a conflict zone could identify their managers as one of their greatest sources of stress. Now I find it considerably easier to understand.
Human beings can tolerate remarkably difficult circumstances when we experience meaning, trust, fairness, belonging and some sense of agency within them. Conversely, even objectively comfortable workplaces can become deeply distressing when people feel powerless, disregarded, treated unfairly or forced to operate inside systems fundamentally at odds with what they believe to be right.
That places an extraordinary responsibility on anyone who chooses to lead. Because when we accept authority over other people, we are accepting more than responsibility for what they produce. We are accepting some responsibility for the conditions under which a substantial part of their life will be lived.
Need help applying these ideas inside your organisation? I work with boards and executive teams to build crisis-ready capability across leadership, culture and systems.