The Unaccountability Machine: When Systems Go Mad (and No One Gets Fired)

Dan Davies—veteran analyst, ex-regulator, and professional myth-buster—has written what might be the most depressingly accurate book of 2024: The Unaccountability Machine: Why Big Systems Make Terrible Decisions—and How the World Lost Its Mind. It is, to put it politely, a masterclass in how large institutions manufacture stupidity while shielding themselves from any consequences. And, even more impressively, they do it without breaking a sweat.

The thesis is simple: institutions, particularly large ones, are not built to solve problems. They are built to survive problems. That distinction matters. Because once an organisation reaches a certain size, it starts to develop defence mechanisms—structures, processes, committees—whose primary job is not to ensure good outcomes but to ensure no one can be blamed for bad ones.

What results is the unaccountability machine: a self-lubricating apparatus that digests responsibility, metabolises it into plausible deniability, and emits a fine mist of corporate jargon.

And yes, you've seen it. Every time a multinational apologises without admitting fault. Every time a government launches a task force on a problem it helped create. Every time a billion-dollar company blames its “internal processes” for what can only be described as a Category 5 ethics collapse.

Davies is less outraged than amused, which is what makes this book sing. He doesn’t scream “burn it all down!” He leans in, pokes the wiring, and says: “Hmm. That’s odd. Why is this system designed to make no one ever feel responsible for anything?”

His concept of “unaccountability sinks” is particularly delicious. These are the departments or roles—compliance, ESG, stakeholder relations—that exist less to solve issues and more to hold them. Like the office microwave that absorbs all smells without ever getting cleaned. They are institutional deodorants: strong on performance, weak on actual hygiene.

This is not theoretical. It's everywhere.

A hedge fund loads up on opaque derivatives. The board is shocked—shocked!—to find risk. A biotech startup flubs a trial, but hey, everyone followed procedure. A health regulator delays a critical approval, but “due process” was followed to the letter. You can almost hear the collective shrug of a thousand bureaucrats: "Mistakes were made. But not by me.”

The book’s brilliance lies in how it reveals that this isn’t malice. It’s engineering. These systems are optimised not for impact, but for resilience to blame. Everyone has dotted their i’s, filed their memos, attended their steering committees. And so nothing happens—and no one gets fired.

At Capital for Cures, I sometimes wonder how long before I accidentally create my own unaccountability machine. A platform is meant to scale, after all. But the moment you start optimising for cover-your-arse instead of clarity, you’ve lost the plot. Innovation dies not with a bang, but with an Outlook invite.

Davies doesn’t offer easy answers—he’s too clever for that. But he does suggest one modest proposal: reconnect actions to consequences. Make ownership traceable. Design feedback loops that sting a little. Ditch the performative consensus rituals and start giving decision-makers enough rope to succeed—or hang themselves. (Metaphorically, of course. HR would never approve.)

Why does this matter? Because in biotech, policy, or healthcare, systems without accountability become systems without outcomes. And the longer we allow decision-making to be outsourced to processes, the more absurd the results. The tragedy isn’t that no one knows what’s going on. It’s that everyone knows—but also knows it’s someone else’s problem.

So yes, read the book. Laugh. Cry. Highlight furiously. And then ask yourself: if something went wrong in your organisation, would anyone take the fall? If the answer is no, congratulations—you may be in the presence of a world-class unaccountability machine. Don’t worry. You’re in good company.

Just don’t be surprised when the next disaster gets a five-slide explainer and a follow-up survey instead of a resignation.