When Genius Failed Ch. 4: When the Market Knows You Must Sell

Investment Background

By early September 1998, LTCM's capital had fallen from $4.7 billion to roughly $2.3 billion. That was already a catastrophe, but technically the fund was alive.

What actually killed it happened over the next two weeks. And that was not a market event.

In early September, Meriwether did something he had no choice about: he began calling Wall Street's major institutions seeking a capital injection.

To let them evaluate the opportunity, he had to disclose the fund's positions.

So within a very short window, Goldman, Merrill, Morgan and several other highly capable trading houses all knew exactly three things:

1. What LTCM held.
2. How large it was.
3. That LTCM had to sell.

The Wall Street Translation

This is the most important chapter in the book, because what it describes does not exist in any textbook risk model.

Textbooks assume prices are exogenous — set by supply and demand, with your trade an insignificant part of it. Under that assumption you can always sell at the market price.

That assumption holds in normal times and fails exactly when you need it.

When the market knows a large seller must liquidate within days, the rational move is not to buy.

The rational move is to sell first.

If I know you must unload $10 billion of Italian government bond spread positions, I short it today. The price falls as you sell, and I buy it back lower.

This is not manipulation and it is not illegal. It is simply rational behavior given known information.

And when every capable institution behaves rationally in the same way, the result is:

Every one of LTCM's positions moved in the worst possible direction for LTCM — and moved before LTCM had even started selling.

A Critical Distinction

Two easily confused things need separating.

Case one: liquidity dries up. Panic, buyers vanish, bid-ask spreads widen. This is passive and treats everyone alike.

Case two: liquidity actively moves away from you. The market knows who you are, what you hold, and what you must do — so prices adjust before you act.

These are not differences of degree.

In case one, your cost is a wider spread. In case two, your counterparty is using your own necessity against you.

What LTCM met in September 1998 was case two.

The distinction matters because case one can be solved by waiting and case two cannot. Waiting is precisely what you do not have — margin calls compel you to act.

The Ultimate Irony of Convergence Arbitrage

Recall the Chapter 1 strategy: buy the cheap thing, short the expensive thing, wait for convergence.

In a normal market, a widening spread means the opportunity got better.

If the on-the-run/off-the-run spread widens from 0.05 to 0.10, the correct response for an unlevered investor is to add. The same convergence now pays twice as much.

This is exactly what LTCM's models said. And the models were right.

But:

At 25x leverage, a widening spread first means a loss; a loss means a margin call; a margin call means you must sell.

So at the moment the opportunity is best, you are forced to do the precise opposite of what the opportunity calls for.

You must sell what you own at its cheapest, and buy back what you are short at its dearest.

Here is the book's central sentence, stated as plainly as possible:

Leverage converts opportunity into threat. The identical price move is a buy signal to the unlevered and a death notice to the levered.

The Rescue

On September 23, the Federal Reserve Bank of New York convened the heads of Wall Street's major banks.

The outcome: fourteen institutions injected roughly $3.6 billion and took 90% of LTCM's equity. The partners' stakes were diluted to nearly nothing.

It must be said clearly that this was not a taxpayer bailout. The Fed put in no money; it supplied a conference room and applied pressure. The money came from private banks — and their reason was entirely self-interested: if LTCM liquidated in a disorderly way, its positions would be dumped onto the market, and those same banks held large amounts of similar positions.

In other words: rescuing LTCM was those banks rescuing their own balance sheets.

The Fed's concern was systemic. LTCM had derivative contracts with nearly every major financial institution, with notional principal exceeding $1 trillion. A disorderly default could have triggered a chain reaction.

This was an early live rehearsal of "too big to fail" — a full decade before 2008.

The Numerical Ending

Item	Value
January 1998 capital	roughly $4.7 billion
Capital remaining at rescue	roughly $400 million
Investor loss	roughly 92%
Rescue injection	roughly $3.6 billion
Fund wind-down completed	early 2000

Meriwether's own net worth fell from hundreds of millions to near zero. Scholes and Merton also lost most of their personal wealth.

Executable Trading Rules

1. Never let yourself enter the state where others know you must act. This is the one rule in the chapter that genuinely matters. For an individual investor there are three main routes into that state: margin accounts, selling naked options, and holding large option positions near expiry. All three convert "choice" into "necessity" at a specific moment.

2. Understand the retail version of the same mechanism. Goldman will not target you; you are too small. But the structure is identical: when your margin account touches the maintenance line, your broker's automated liquidation sells your holdings at the market's worst prices, and it will not wait for you. Forced liquidation never happens on a good day.

3. At the moment you open a position, ask: "if this goes against me, who can force me to close it?" If the answer is "nobody," you are safe. If any name appears — broker, clearinghouse, lender — reconsider the size.

4. Write down, in advance, your response to a widening spread. Unlevered, the right response is add or hold. Levered, you have no choice but to sell. That difference decides whether you are a buyer or a seller on the day that matters most.

5. Do not believe "I will exit before things get bad." LTCM had the world's best risk monitoring, real-time position data, and Nobel laureates watching the screens. They could not exit. Because exiting was itself the thing driving prices down.

Relevance to a Retirement Portfolio

This chapter explains why we repeat something apparently boring across this entire site:

Preserve your ability to do nothing on the worst day.

A retirement portfolio of low-cost index funds and cash fell a long way in 2008. Holding it was painful.

But not one person holding it was forced to sell in 2008.

Which is the only reason they fully recovered afterward.

LTCM's strategies were largely correct — Chapter 5 will prove this with data. The spreads did converge. The fund did not survive to see it.

For your retirement portfolio, this chapter translates into one sentence:

Your objective is not to maximize returns. It is to ensure that on no day can anyone force you to sell your future at the wrong price.

And achieving that requires no intellectual advantage whatsoever. It requires only that you give up leverage.