Why Good People Participate in Harmful Institutional Systems

The puzzle isn't why cruel people do cruel things. That part isn't hard to explain. The harder question — the one that actually matters for understanding how institutional harm sustains itself — is why kind, conscientious, genuinely decent people keep it running, often for years, often while privately believing that what they're participating in is wrong.

This is the pattern I keep coming back to. Not the villain. The idealist who eventually became an enforcer. The bystander who became complicit. The person who walked in with high hopes and walked out, decades later, having helped an institution do things they would never have chosen freely on their own. If you've spent any time inside large organizations — hospitals, corporations, universities, government agencies, religious institutions — you've likely met this person. You may have been this person.

Understanding why it happens is, I think, one of the more important structural questions we can ask about how institutions actually work.

The Frame That Fails Us

The instinctive explanation for institutional harm is personnel. Find the bad actors, remove them, and the system heals. This is almost always wrong — not because bad actors don't exist, but because it misidentifies what's doing most of the work.

In 1963, Stanley Milgram conducted what became one of the most replicated studies in social psychology. Ordinary Americans — recruited from newspaper ads, representing a wide range of occupations and backgrounds — were told they were participating in a learning study. When a test subject in the next room gave a wrong answer, they were instructed to administer an electric shock. The voltage increased with each mistake, up to a labeled maximum of 450 volts. There was no actual shock — the test subject was an actor — but the participants didn't know that. Sixty-five percent of participants administered what they believed to be the maximum voltage, with many visibly distressed throughout. The defining variable wasn't cruelty. It was the presence of an authority figure in a lab coat saying "please continue."

Hannah Arendt, observing Adolf Eichmann's trial in 1961, arrived at a similar insight from a different direction. Eichmann wasn't a sadist. He was, in Arendt's word, "thoughtless" — not stupid, but systematically disengaged from the human consequences of his administrative decisions. She called this "the banality of evil," and the phrase landed hard because it described something most people already sensed but hadn't quite named: that the most durable institutional harms are not sustained by monsters but by conscientious bureaucrats doing their jobs.

Organizational research consistently finds that role structure and normative pressure are stronger predictors of ethical compliance than individual moral character — meaning the problem is in the system's design, not the employee's soul. This is deeply uncomfortable, because if the problem were simply bad people, the fix would be simple. It's not.

Five Mechanisms That Capture Good People

These are the mechanisms I think are doing most of the structural work. They're not exotic or rare. They operate in ordinary workplaces, in institutions you'd recognize, in people you'd admire.

Role Absorption

When you join an institution, you accept a role — and the role comes pre-loaded with a way of seeing. A vocabulary. A set of things that count as "the work" and a set of things that count as someone else's concern. Over time, the role substitutes for your own judgment. The question shifts from "is this right?" to "is this how we do it?" That shift is almost imperceptible from the inside, and it happens fast. Most organizational socialization research suggests the core identity shift happens within the first year of a new position.

What matters is that this doesn't feel like coercion. It feels like growing up. Getting competent. Learning to think like a professional in the field. There's genuine satisfaction in it — the satisfaction of mastery, of belonging, of being trusted with real responsibility. The problem is that every institution has evolved blind spots it needs to protect, and when you learn to see the way the institution sees, you stop seeing what the institution can't afford to see. The blind spots become invisible by design.

I've seen this operate in highly educated, morally serious people. The role doesn't override their values — it repositions those values so they're serving the institution's ends rather than their own.

Incremental Escalation

Nobody starts at the extreme. You start somewhere defensible: sign off on a report that overstates results by a small margin. Don't raise the quality concern this quarter because the launch timeline is tight. Look the other way on one customer complaint because the relationship matters. Each step is small. Each step can be justified by reference to the previous one.

By the time the harm is obvious, you are years into it. And here is the trap: the only way to acknowledge the problem is to acknowledge your own participation in it. For most people, that's a cost they can't pay. Classic foot-in-the-door research in moral psychology shows that small initial commitments dramatically increase the probability of larger later ones. The escalation doesn't feel like escalation from inside it. It feels like consistency — and consistency is a virtue.

This is how Theranos employees kept working. How Enron accountants kept processing. How Boeing engineers kept signing off. Not because they were suddenly different people, but because each step had been normalized by the one before it.

Diffused Responsibility

In complex institutions, no single person does the whole harm. The engineer designs the component. The manager approves the budget. The lawyer reviews the contract. The executive signs the filing. The regulator stamps the approval. Each person in the chain did their piece. Each can honestly say, in a narrow sense, that they didn't "do it."

This is what I'd call the architecture of moral invisibility. Organizational psychologists call it diffusion of responsibility — the same mechanism that makes bystanders less likely to help an accident victim when more witnesses are present. In Stanley Milgram's follow-up studies, when the shock administration was split across multiple participants, compliance rates increased. The more people who could have objected, the less any one person felt they needed to. Institutional complexity manufactures this condition at scale, often without anyone intending to do so.

The result is that harm requiring the coordinated action of hundreds of people can be sustained indefinitely without any of those people feeling individually responsible for it.

Institutional Legitimacy as Moral Cover

Legitimacy is more powerful than most people recognize. When an institution has a long history, respected credentials, a convincing public purpose, and visible social approval, the people inside it experience something like moral permission. The institution has already been judged. Already been approved. Your job is to execute, not to re-adjudicate the whole enterprise from scratch every morning.

This is how universities can systematically damage student mental health while sincerely describing themselves as student-centered. How pharmaceutical companies can promote addictive drugs while genuinely talking about patient welfare. How financial firms can extract value from clients while calling themselves advisors. The language of purpose provides air cover for the mechanics of extraction — and in my experience, the people delivering the harm have often genuinely internalized the purpose language. They're not lying when they say they believe in the mission. They believe in the stated mission. The actual mission is operating several floors below their awareness.

Legitimacy functions as moral outsourcing. You don't have to evaluate the institution every day because the institution has already been evaluated by people you respect. This is efficient in organizations that are genuinely good. It's ruinous in organizations that have learned to wear the appearance of goodness.

The Cost of Exit

Even when someone inside a harmful institution clearly sees the problem, leaving isn't free. Walking out means surrendering salary, benefits, pension, professional network, professional identity, and often the only career they know. For someone in mid-career, these costs can be enormous in practice — not abstract enormous, but "my kids' tuition" enormous, "my health insurance" enormous.

There's also the asymmetry of the bet. If you leave and the institution turns out to be fine, you sacrificed everything for a stand that turned out to be misplaced. Certain personal cost against uncertain institutional harm — most people resolve that bet in favor of staying. And cognitive dissonance does the rest: the longer you stay, the more your mind constructs justifications for the decision to stay. Research on organizational exit consistently finds that employees who recognize ethical problems stay significantly longer than they anticipated when they first joined the organization.

Exit is the mechanism the institution depends on being too costly to use.

Comparing the Explanations

The structural explanation is the uncomfortable one, because it means the problem is in the institution's design — and changing an institution's design requires confronting the people whose interests are served by the current arrangement.

What Happens When Good People Notice

There's a specific moment that interests me — the moment someone inside a harmful institution clearly sees what the institution is doing. Not suspects. Sees.

What happens next is revealing. The most common response is not whistleblowing or resignation. It's compartmentalization. The person creates a mental separation between "the work I do" and "what the institution does." This is not cynicism exactly. It's closer to a survival mechanism — a way of staying functional while sitting with a truth that has nowhere to go.

Some people try to change things from the inside. They raise concerns, form committees, write careful memos. In my observation, this mostly doesn't work — not because these people aren't smart or committed, but because mature institutions have evolved defense mechanisms specifically calibrated to neutralize internal challenges. Concerns get heard. Studies get commissioned. Task forces get formed. The energy of the challenge gets absorbed and redirected into the institution's own processes, and the system continues. The people who raised the concerns often end up feeling like the problem was addressed, when in fact it was managed.

The people who actually change things are usually not insiders working through proper channels. They tend to be people who eventually chose exit and external pressure — whistleblowers, investigative journalists, researchers with no institutional stake in the outcome, regulators acting from outside. The pattern holds across industries and across decades. Diane Vaughan's research on the Challenger disaster showed that engineers inside NASA had raised concerns about the O-rings before the launch. The concerns were absorbed by the institution's risk-normalization machinery. It took a catastrophe and an external investigation to name what had actually happened.

What this tells me is that the good people inside a harmful institution are genuinely valuable to it — not primarily as agents of change, but as proof of credibility. Their presence signals that the institution is not purely extractive. That signal is worth a great deal. The institution, consciously or not, depends on it.

The Pattern, Plainly Stated

What I think is happening in these cases is not a failure of character. It's a success of design. Institutions that survive and grow across decades get good at capturing and converting the moral energy of the people inside them. Good people don't sustain harmful systems despite their goodness — in many cases, they sustain them because of it, because their commitment, their work ethic, their genuine belief in the mission is exactly what the institution needs to keep the surface credible while the extraction runs below.

There's a reason institutional scandals almost always end with the conclusion that a few bad apples were responsible. In my view, that conclusion is almost always the institution protecting itself. The real story is structural — and the structural story is harder to tell, harder to prosecute, and far harder to fix. It doesn't have a villain. It has a design.

If you want to understand why an institution causes harm, studying the character of the people inside it will tell you very little. Study the structure instead. Who benefits from current arrangements. What behaviors the institution actually rewards versus what it says it values. What happens to the people who raise serious questions. Those facts are usually more available than we think. And more revealing than any character assessment could ever be.

The question worth sitting with: have you noticed how rarely "we need to change the structure" is the conclusion when harm comes to light? And have you wondered why that is?

[Related: How Institutional Capture Spreads | The Pattern of Normalization of Deviance]

Last updated: 2026-06-02