Every funding crisis I have worked through began the same way: with a senior person, in a calm room, pointing at a ratio and saying that the institution had plenty of liquidity. The ratio was accurate. It was also nearly irrelevant, because liquidity is not a stock you hold. It is a behaviour other people exhibit — the behaviour of continuing to lend to you, roll your funding, and accept your collateral — and behaviour is exactly the thing that changes in a crisis. [author assertion]
The number on the report measures a quiet day. It counts assets that could be sold and facilities that could be drawn, at prices and on terms observed when nobody needed them. But the moment liquidity is actually required, the measurement's assumptions go to be tested all at once, by everyone, in the same direction. Liquidity is the only asset whose quantity depends on how many people are trying to count it at the same time. [author assertion]
Watch a funding market seize and you will see behaviour, not arithmetic. Lenders do not first reprice a troubled borrower; they shorten. Thirty-day money becomes overnight money. Then overnight money requires a phone call, and the phone call goes unanswered. The ladder of tenors collapses from the long end, quietly, over a few days — and each lender's caution is perfectly rational, which is exactly why the aggregate is catastrophic. [author assertion]
Haircuts tell the same story. The asset in the vault did not change on the morning its haircut doubled. What changed was the crowd's willingness, and willingness is contagious in both directions. A risk framework that treats haircuts as parameters of the asset, rather than as opinions held by a crowd, will be surprised at the precise moment surprise is most expensive. [author assertion]
The institutions I saw come through funding stress in the best shape were not, in my experience, the ones with the most impressive buffers on paper. They were the ones whose funding was boring in a specific way: diversified by behaviour, not merely by name. Their lenders were different kinds of institutions, in different jurisdictions, running money for different reasons — so that no single story could make them all run at once. Diversity of names with unity of motive is not diversity. It is a crowd wearing name badges. [author assertion]
So my practical rule for any treasury I advise: stress the relationship, not the ratio. For each material source of funding, ask who is on the other side, what would make them leave, and whether the things that would make them leave are the same things that would make everyone else leave. If the answers rhyme, the liquidity is thinner than the number says. [author assertion]
Liquidity, properly understood, is a promise that other people have made to you implicitly, under conditions they have not specified, revocable without notice. Count it the way you would count any such promise: as something real, valuable, and not yours. [author assertion]
