Toxic bosses: why do companies protect managers who make working life hell?

When Andrea finally resigned, her colleagues told her the truth they had always known about her boss. Some apologised. Others admitted they had watched what happened to her and said nothing. One offered the sentence I have been unable to forget: “You’ve been an emotional stab vest for us. You protected us.”

Andrea was a senior legal executive with decades of credibility when she began reporting to a powerful new colleague. Within months, something was wrong. Expectations shifted without warning and feedback contradicted itself. He spoke over her in public and undermined her in private. Slowly she began questioning not just him but herself, her competence, even her grip on reality.

What disturbed Andrea most was the gradual discovery that everybody already knew. Colleagues confessed that when they heard this executive was returning to the organisation, they had cried. When Andrea approached HR, a senior executive told her: “People work longer at McDonald’s than they work for him.” That was acknowledgment, not discovery. And there is something brutally revealing in the stab-vest description. Andrea had become the organisation’s psychological shock absorber. While she took the blows, everybody around her was safer.

Stories like Andrea’s are usually told as stories about monsters: the bullying boss, the narcissistic founder who finally, spectacularly, falls. I have come to think of such framing as the least useful way to tell them. Instead, I’m interested in how such people keep acquiring power and why, once their behaviour is visible to almost everyone, nobody stops them.

I have encountered power through three lenses. For 20 years, I was an executive-level headhunter, spending thousands of hours assessing the people organisations were preparing to entrust with authority. I then held that authority myself, as a global chief marketing officer in New York. Now, as a psychotherapist, I meet power from the opposite side, through people trying to understand what happened to them when somebody wielded it badly. The industries differ; the consequences rarely do. Brilliant people become convinced they are incompetent. Experienced professionals stop trusting their own judgment. Good leadership enlarges people. Bad leadership, at its darkest, persuades a person, slowly, to disappear.

It never begins that way. The leaders who do the most psychological damage rarely arrive as tyrants. They arrive as answers, wrapped in charm and certainty, mirroring your values back to you. That first meeting can feel like the best of your professional life. Clinically, this can create a rapid sense of attachment, in which trust and emotional closeness develop before the relationship has had time to be tested. What follows – withdrawing warmth and then restoring it just often enough – is the logic of the slot machine: intermittent reward keeps people hopeful and vigilant, working ever harder to recover the golden beginning. By the time the storm breaks, the job has been accepted and the mortgage signed – leaving feels unthinkable.

Some of this starts at selection. Consider the language of the senior job specification: bold, decisive, resilient, charismatic, comfortable under pressure, possessed of “executive presence”. There is nothing wrong with any of it. The problem is that a healthy leader and a dangerous one can look almost identical at interview. One candidate is decisive because they have experience and a secure enough ego to make the difficult call; another because doubt itself is intolerable. One is resilient through self-regulation, another through emotional detachment. Both interview beautifully. Corporate language launders the difference: low empathy is repackaged as decisiveness, ruthlessness as “high standards”, emotional detachment as resilience, manipulation as influence.

In 2015, a meta-analysis led by the psychologist Emily Grijalva found that narcissism meaningfully predicts leadership emergence, which is to say getting the job, but has almost no relationship with leadership effectiveness. People high in the so-called dark triad traits – narcissism, machiavellianism and psychopathy – are disproportionately skilled at impression management, and although psychopathic traits appear in roughly 1% of the general population, studies of corporate samples suggest markedly higher concentrations among business leaders.

As a therapist, I am also interested in what forms such leaders. Research published in the Leadership Quarterly found an association between supervisors’ experiences of family undermining during childhood and abusive supervisory behaviour in adulthood, particularly among those with lower self-control. This demands careful interpretation. Trauma in itself does not create toxic leaders and disqualifies nobody from leadership. Many remarkable leaders have transformed painful experience into empathy and wisdom.

The distinction that matters is between trauma and unprocessed trauma. There is a profound difference between a leader who has made sense of their lived experience and one whose unexamined wounds continue to seep into the way they use power. The child who survived chaos by monitoring everything can become the executive who must control everything. Somebody who once felt profoundly powerless may spend an adult lifetime making sure they never feel powerless again.

The survival mechanisms do not necessarily disappear. They can professionalise. Hypervigilance becomes “strategic oversight”. The inability to trust becomes “strong leadership”. Emotional suppression becomes “composure under pressure”. What once protected a person from vulnerability can become, under the authority of office, the very thing that makes everybody around them unsafe.

The psychoanalyst and organisational psychologist Manfred FR Kets de Vries has spent decades examining the psychological forces beneath executive ambition, and in conversations I had with him about narcissism, his warning about what happens when personality meets authority has always stayed with me. “When people get into a situation of power, unfortunately, narcissism is a very heavy drug and many people cannot handle it,” he said in a 2025 interview. The corporate world can compound that risk because it often rewards the very behaviours it produces: intensity, certainty, relentless drive and control.

But explanation is not exoneration, and an explanation of the leader must never become an alibi for the institution. The facts of each case vary, but what is most striking about the public cases is not what the leader did but how many people knew. The CBI’s implosion under misconduct allegations, the Ted Baker staff petition against their founder’s culture of “forced hugging”, Harvey Weinstein as Hollywood’s open secret: in account after account, the recurring detail was that none of it was hidden. The Ethics & Compliance Initiative reports that about three-quarters of employees who witness wrongdoing do report it, either internally or externally. The signals arrive, yet too often they fail to prompt meaningful action, particularly when the person involved is judged too valuable to lose.

Around a destructive leader grows what I think of as an architecture of accommodation: colleagues who privately disapprove, boards that soften the language, investors who admire the numbers, and employees who decide, entirely understandably, that their mortgage matters more than becoming the next target. Nobody has to intend to protect abuse. Each person simply has to make the safest decision available to them.

Inside a hierarchy, these forces compound. We doubt what we are seeing because nobody else appears alarmed, we assume somebody senior must know more than we do, and we watch what happens to the first person who speaks before quietly deciding not to become the second.

The silence is expensive. When MIT Sloan researchers analysed 34m employee profiles, toxic culture proved 10 times more powerful than pay in predicting who quit. Britain’s Health and Safety Executive identifies the way managers manage staff as a major cause of work-related stress, and Stanford’s Jeffrey Pfeffer, studying workplace mortality, concluded that harmful management practices are “literally killing people”. We banned smoking in offices, so why do we still tolerate this?

Curiously, we govern almost everything except behaviour. A serious company has audit committees, risk registers, compliance functions and external auditors, because financial integrity is never entrusted to goodwill. Yet when the issue is the conduct of the most powerful person in the building, we become oddly informal. It’s a “leadership style”, a “personality clash”, a HR matter, something sensitive to be mediated quietly. And HR is structurally the weakest player. Research I conducted across the media industry in 2022 found that HR ranked as the least respected function by senior executives, which means the people charged with guarding behaviour are marginalised precisely where power most needs guarding.

This is the point at which personality becomes organisational. One person’s personality can become policy. One person’s psychology can become culture. One person’s appetite can become the company’s strategy. Personality, when combined with authority, can become a system.

So what would behavioural governance look like in practice? Boards could independently test whether people can safely disagree with senior leaders. They could look for patterns in attrition rather than treating every departure as a separate event, connect complaints, settlements and confidentiality agreements rather than letting organisational memory reset with each exit, and assess a leader at three, six and 12 months in, once the interview performance is over and the experience of working beneath that person has begun.

Above all, such mechanisms must be independent, because asking a function embedded within a hierarchy to investigate the most powerful person in that hierarchy is an obvious conflict. If we insist on independent governance for finance, why do we accept its absence for people?

The law, at least, is beginning to evolve. Since October 2025, new NDAs in England and Wales cannot prevent victims, or those who reasonably believe they are victims, from making specified disclosures about criminal conduct. The Employment Rights Act 2025 will go further: when its NDA provisions take effect, contractual clauses will be void insofar as they prevent workers speaking about harassment or discrimination, or about how an employer handled their complaint.

Zelda Perkins, the former Weinstein assistant whose Can’t Buy My Silence campaign has been at the forefront of this reform, described NDAs to me as sitting “at the foundation of that architecture of silence”, concealing patterns of behaviour and preventing organisations from recognising the risk. It is a striking description because secrecy does more than protect one individual. Left unchallenged, one person’s pathology can become an organisation’s dysfunction.

Workplace bullying itself, though, is not a stand-alone criminal offence, and Britain still has no specific law against it. Legislation is beginning to tackle the secrecy surrounding workplace abuse. But protection against bullying still lags behind.

Toxic leadership survives through repetition. One generation learns to tolerate it and quietly teaches the next how to survive it: keep your head down, toughen up, don’t become difficult, protect your income, guard your reputation, learn to manage the boss. That inheritance can be refused by the board member who asks the harder question, the colleague who declines to join the whispering afterwards, the investor who asks how the numbers were achieved, the organisation that protects the person telling the truth rather than the person holding the power.

The people who have benefited from our silence are counting on us to keep it that way. We should give them reason to think again. Through structured governance, yes, but more than anything by treating our colleagues with something far less onerous – simple humanity.