Cameron. May. Johnson. Truss — who lasted forty-nine days. Sunak. Starmer. And now Burnham. That is seven British prime ministers in ten years, and not one of them served a full term — the fastest turnover of national leaders Britain has seen in almost two hundred years.

There was no coup, no revolution, no tanks in the street. Each one was removed, quietly and legally, by the same simple mechanism — a mechanism doing exactly what it was built to do. And it all comes down to one question every democracy has to answer: how easy should it be to remove the person in charge?

Two years ago Keir Starmer won a landslide — a majority of 174 seats, one of the biggest in modern British history. This June he was gone: no election, not even a formal vote to remove him. His own MPs simply stopped supporting him after a run of bad results, and he walked away. His replacement, Andy Burnham, steps into the job this month — again, with no general election; the public doesn't get a say until the next one, years from now.

It would be easy to conclude that British politicians are just uniquely bad at their jobs. But the list rules that out: five of the seven were Conservatives, one was Labour. The churn hit left and right, landslide winners and narrow ones, the scandal-ridden and the simply unlucky, all the same. When a problem shows up no matter who is in charge, you're not looking at a people problem — you're looking at a design problem.

Over the exact same ten years that Britain burned through seven prime ministers, Germany had roughly two chancellors. Same kind of democracy, wildly different stability — the politicians aren't the variable, the rules are. And underneath those rules sits a single, genuinely hard trade-off: you want a leader you can remove when they're bad, but you also want one stable enough to actually govern. Those two wishes pull in opposite directions.

British voters don't elect a prime minister — they elect a local MP, and the biggest party forms a government with its leader as PM. So the person who can fire the PM isn't the public; it's their own party, a few hundred MPs. Your real electorate isn't sixty million people — it's the room down the hall. And to remove a leader, those MPs only have to agree on one thing: that the leader should go. They don't have to agree on who replaces them, or on any plan — and that's a very low bar, because tearing down is far easier than agreeing what to build. So the trigger is fear (bad polls, a lost local election), not the arrival of a better alternative.

Then first-past-the-post pours on the fuel. Because whoever gets the most votes in each district wins it, a small drop in a party's popularity can mean an enormous loss of seats — so when the polls dip, every ordinary MP sees their own seat disappearing, and the fastest fix is to swap the leader and hope for a bounce. Bad polls create fear, fear takes out the leader, the new leader inherits less authority and a shorter fuse, and the loop spins faster. None of the three is corrupt; each was reasonable alone. Together they build a machine that optimizes for one thing — replacing the leader — on a timescale of months, when governing takes years.

There are really only three levers, each landing in a different place on that dial. The first is the term: in a presidential system like the US, the leader is elected for a fixed term and can only be removed for serious wrongdoing, through impeachment. The gain is stability — a horizon of years; the cost is that if you elect someone reckless or out of their depth, you're stuck. In a fixed-term system there's no quick exit from a leader who fails as fast as Truss did.

The second keeps the power to remove but changes how. Between the wars, Germany's parliament could throw out a chancellor with a simple no-confidence vote — and more than twenty governments collapsed in fourteen years, a chaos that helped open the door to the Nazis. So Germany added one rule: parliament can remove the chancellor only by electing a replacement at the same moment. You can't just tear down; you must agree who builds next — which is why this weapon has been used successfully only once in seventy-five years. The cost: a fragmented opposition can leave an unpopular government stuck. The third lever changes the voting system underneath: switch first-past-the-post for proportional representation, where seats match vote share, and a poll dip stops being a wipeout — so the reflex to knife the leader loses its fuel. The cost is coalitions, and their own instability. None of these is free, and none is obviously right — every setting buys stability by giving something up.

Here's a detail that makes it sharper. Britain once tried to add a little stability: in 2011 it passed the Fixed-term Parliaments Act, fixing elections to a set schedule — a brake on one kind of political impulsiveness. And then, in 2022, it repealed it, taking even that brake back off. Because there's no obviously correct answer here. A leader you can remove the moment they fail is accountable — but fragile. A leader you can't remove is stable — but dangerous when they're wrong. Every country is quietly parked somewhere on that dial, usually without ever having chosen the spot on purpose. And this isn't really about prime ministers: any company, board, or team answers the same question — how easy is it to remove the person at the top? Make it too hard and you're trapped with your mistakes; too easy and no one can plan past next week. So where would you set the dial — and do you actually know where yours is set right now?