When the Exits Vanish: How Funds Run Out of Liquidity - Part 1

Funds rarely die from being wrong. They die from being unable to wait around long enough to be proven right. That distinction sits at the center of almost every famous blowup. Solvency is a question about whether the trade was correct while liquidity is a question about whether you survive long enough to collect. A fund can be right about the destination and still get carried out because the thing that kills it is not the final price; it's everything that happens on the way there. The market can stay illiquid longer than a levered fund can stay funded.

What follows is a walk through the machinery that turns a paper loss into a forced unwind. The cases differ in their triggers, but they rhyme. In every one, the exit that everyone assumed would be there simply vanished.

Two Kinds of Liquidity

Before the cases, one framework that the rest of this hangs on. There are two distinct kinds of liquidity, and confusing them is where a lot of risk management quietly goes wrong.

• Market liquidity is the ability to sell an asset without moving its price against you. A position is liquid when you can exit at something close to the last printed price. It becomes illiquid the moment your selling itself starts pushing the price down.
• Funding liquidity is the ability to borrow, or to meet a margin call, so that you don't have to sell in the first place. It's the financing that lets you hold a position through a drawdown rather than liquidate into it.

The danger is rarely either one alone. It's the loop between them. A funding squeeze forces selling. Selling drains market liquidity and pushes prices lower. Lower prices trigger fresh margin calls, which force more selling. The spiral feeds itself, and it runs faster than most risk models assume because both kinds of liquidity tend to dry up at the same time and for the same reason. The exit narrows precisely when the largest number of people are trying to use it.

Hold that loop in mind. Every case below is a different way of setting it off.

The Leverage Spiral: LTCM

Long-Term Capital Management (LTCM) is the canonical story, and it is canonical for a reason. The fund ran relative value and convergence trades, positions that were individually sensible and built on the patterns that small pricing gaps between closely related instruments tend to narrow over time. On any single trade, the edge was thin but the logic was sound. The problem was what sat on top of those trades. To make thin edges meaningful, LTCM stacked enormous leverage, financed across a web of counterparties. That leverage did something subtle: it compressed the amount of time the fund had to be right.

In 1998, against the backdrop of Russia's default and a broad flight to quality, spreads did the opposite of converging. They widened. Positions that were supposed to be diversified moved together, because in a stress event the common factor is not the instrument, it's the desperation of the people holding it. The margin calls arrived faster than the positions could be unwound, and unwinding them at all meant selling into markets that had already thinned out. The fund was not obviously insolvent. The trades, given enough time, might well have converged. LTCM just didn't have the funding to hold them long enough, and the eventual recapitalization by a consortium of banks was an admission that an orderly exit no longer existed at any price the fund could survive.

The lesson worth pulling out is not "leverage is dangerous." It's more specific than that. Leverage doesn't just amplify losses; it shortens your runway. It converts a question of whether you're right into a question of whether you're right soon enough.

The Crowded Trade: When Everyone Owns the Same Exit

A position can look perfectly liquid on your own screen and be deeply illiquid in reality, because liquidity is a function of who else is holding the same thing. Amaranth Advisors offers the concentrated version of this. The fund built a very large, very directional set of natural gas spread positions - big enough that it had effectively become the market in certain contracts. Being the market is fine until you need to leave it. When the trades moved against the fund in 2006, there was no one of comparable size to take the other side, and the act of trying to exit moved prices further against the remaining position. The exit wasn't closed by a panic. It was closed by the fund's own footprint.

The August 2007 "quant quake" shows the distributed version of the same problem. Many quantitative funds, working independently, had converged on overlapping factor positions. None of them could see the others' books, so the crowding was invisible from any single desk. When one large player began deleveraging, its selling pushed those shared factors against everyone else holding them, which triggered more deleveraging, which pushed prices further. Funds that had never spoken to each other were suddenly forced sellers of the same names at the same moment.

The pattern in both cases is the same. Crowding is a hidden liability. It doesn't show up in your own risk report because your risk report cannot see who's standing next to you. The trade is liquid right up until the instant everyone reaches for the identical exit, and then it isn't.

Come back tomorrow for Part 2 of this topic!