The Northern Rock bank run wasn't irrational. It was the only rational response available.

Depositor behavior was completely logical. What was irrational was the bank's business model.

Queues formed at a bank's door and the press called it irrational panic. They inverted the adjective. The behavior in the queue was the most rational thing in the plot. The irrational part was on the balance sheet.

A bank does a strange and necessary thing: it takes money that can be withdrawn at any moment and lends it out for very long terms. You deposit today and can withdraw tomorrow; your mortgage lasts thirty years. That transformation of maturity is the heart of banking.

The trick only works because of one assumption: not everyone will want their money back at the same time. The bank bets on dispersion. As long as withdrawal requests arrive in a trickle, it pays from the day's cash and carries on. It is an equilibrium of trust, not of solvency.

Northern Rock took this trick to an extreme and in a worse version. Instead of funding its mortgages with deposits from many different people, it funded itself in the market, taking short-term money from other institutions to lend out for thirty years.

Notice the fragility of the design. The deposits of thousands of people are, on average, stable: each decides on their own, and the decisions cancel out. Market funding is a herd: a few large actors who decide together and flee together.

When the credit market froze, that short-term funding simply did not renew. The source dried up overnight. The bank still had good long-term assets and no way to pay its short-term bills. Solvent on paper, out of cash in practice.

Now put yourself in the queue. You are an ordinary depositor. You know the bank depends on funding that has stopped. You know that if many people withdraw, there won't be enough for the last ones. What is the rational move? Withdraw first. Running is the right answer, not the error.

This is the point almost everyone gets wrong: in a bank run, individual panic is collectively destructive and individually correct. There is no virtue in being the last polite person in line at a bank with no cash. Each person's rationality produces everyone's ruin.

The technical name for this is maturity mismatch: a long, illiquid asset funded by short, volatile liabilities. It is the oldest and most recurrent fragility in finance. No matter the century or the instrument. When the maturity of what you owe is shorter than that of what you hold, you live on a tightrope.

There is an institution invented precisely to defeat this logic: deposit insurance. The idea is simple and elegant. If the state guarantees your money, you have no reason to run. And if no one runs, the prophecy does not fulfill itself. The guarantee works best when no one needs to use it.

The run happened because that guarantee, in this case, was ambiguous: it covered part, not all, and no one was sure of the details in the moment of fear. And an ambiguous guarantee is worse than none, because it leaves room for doubt, and doubt is the fuel of the queue.

The lesson for those who design systems: either the safety net is credible and total enough to make the run unnecessary, or it becomes just one more uncertainty in the middle of panic. A half-guarantee calms no one. It invites everyone to test whether it applies to them.

For the investor, the signal is broader than one British bank. The question to ask about any institution is: how does it fund itself? Whoever relies on money that must be constantly renewed is betting the window will never close. But windows close, and they close precisely when you need them most.

And there is a cruel regularity: the cheapest funding is almost always the most volatile. Short-term money charges less interest because it can flee at any moment. You save in good times exactly what will kill you in bad. Cheap in funding tends to be expensive in risk.

Note the sequence: a long asset funded by short money; the short money flees; the bank runs out of cash despite being solvent; the rational depositor runs; the run confirms the collapse. Each link is logical. It was the design that armed the tragedy, not the emotion of the crowd.

The rule of this month: do not ask whether an institution has good assets. Ask whether it can survive the day its funding disappears. Solvency is about what you hold. Survival is about how you fund it. The second kills first.

Northern Rock was not the victim of irrational people. It was the victim of rational people reacting to an irrational model. And that will be the shape of nearly every collapse to come: fragile structures punished by perfectly logical behavior.

Mark the queue not as hysteria, but as collective intelligence doing its cruel job: revealing, in a few hours, a fragility the bank hid for years.

Leo Bentier