"History doesn't repeat… but it rhymes." — Mark Twain
◉ THE PRESENT
The redemption gate hit at 4:47 p.m. yesterday, after the close. One of the top five multi-strategy hedge funds — $42 billion under management, the kind of place that prints money in any tape — told its investors they could not have their money back. The fund had been short the basis between Treasury futures and cash bonds, levered fifty to one through the repo market, and the trade had moved against them three sessions in a row. By the time the wires picked it up, the VIX was at 31.4 and S&P futures were down another two percent in the overnight.
VIX 31.4 | S&P -3.1% | Repo 6.2% | 10Y 4.62%
The traders on the desks did not need to look up the echo. Most of them had lived through it once already.
◉ THE ECHO — SEPTEMBER 23, 1998
Fourteen men in a room on Liberty Street.
It was a Wednesday evening in lower Manhattan, and William McDonough, the president of the New York Federal Reserve, had called the heads of every major bank on Wall Street into his wood-paneled conference room at seven o'clock. He did not have a plan. He had a problem. A hedge fund in Greenwich, Connecticut had just told him it was about to take the global financial system down with it, and he had until the Tokyo open to figure out what to do about it.
The hedge fund was Long-Term Capital Management. It had been the smartest place on Wall Street since the day it opened in 1994. John Meriwether, the legendary Salomon bond trader, ran the floor. Myron Scholes and Robert Merton, who had won the Nobel Prize the year before for the option pricing formula every trader on earth used, sat in the partners' offices. Their trades were small in any one place but they ran them at twenty-five to one leverage and they made over twenty percent every year for four straight years like clockwork. By January they had $4.7 billion in capital, $125 billion in assets, and a notional derivatives book pushing $1.25 trillion. Trillion with a T.
Then Russia defaulted on August 17th. The ruble collapsed. Spreads on every kind of credit blew out at the same time, in the same direction, on the same Tuesday morning, and the models that LTCM had built — the models that assumed those spreads moved independently — told the partners that they had just lost forty percent. They had not lost forty percent. They had lost almost everything. By the second week of September the capital was down to $600 million against the same $125 billion book, and Bear Stearns was refusing to clear their trades.
McDonough went around the room. He asked them, politely, to put in $3.6 billion of their own money and take the fund over. If they did not, every man at the table would be on the other side of LTCM's positions in a forced fire sale on Monday morning, and they would all blow up together. The deal was done by midnight. Bear Stearns refused to participate. The other fourteen agreed.
Twenty-eight years later, the gate at the pod shop yesterday looked identical.
◉ THE RHYME — WHAT'S IDENTICAL
When everyone runs the same trade with the same model on the same leverage, the rhyme is the math. The model breaks the same way every time.
◉ THE DIVERGENCE — WHAT'S DIFFERENT THIS TIME
The differences matter. They are where the edge lives.
LTCM was one fund. You could draw a circle around the problem and put fourteen men in a room and solve it in five hours. The basis trade today is not one fund. It is dozens of funds running the same Treasury cash-versus-futures arbitrage through the same prime brokers on the same repo desks. The problem is the market structure itself, and you cannot put market structure into a conference room.
The Fed has the Standing Repo Facility now, which did not exist in 1998. It can drop liquidity into the dealer system at five in the afternoon with no congressional permission and no press release. McDonough had to organize a private rescue because the Federal Reserve Act gave him no clean tool. Powell has the tool. The question is whether he uses it before the next gate hits.
The Treasury market is $28 trillion today versus $3.7 trillion in 1998. The dealer balance sheet that has to warehouse the paper has not grown anywhere close to that pace. When ten-year auctions tail by four basis points the way they did three weeks ago, that is the market telling you the buyer of last resort is missing and the plumbing is thinner than the size of the system requires.
Inflation in 1998 was 1.6 percent and falling. Inflation today is 3.4 percent and stuck. Greenspan could cut three times in seven weeks without losing his anti-inflation credibility. The current Fed cannot. Whatever rescue this market gets, it will arrive later and smaller than the one McDonough's people delivered.
◉ THE RECKONING — WHAT HAPPENS NEXT
The McDonough rescue worked. The fourteen banks took over LTCM, wound the book down slowly over a year, and lost less than they would have lost in a fire sale on Monday morning. But the rescue did not stop the damage to the broader market. The S&P 500 had peaked at 1,186 on July 17, 1998. By the night of the McDonough meeting on September 23rd it was at 1,066, already down ten percent. It kept falling for two more weeks, hitting 959 on October 8th, down nineteen percent from the peak.
The Fed cut twenty-five basis points on September 29th at the regular meeting. The market did not believe it. Greenspan cut again on October 15th, a Thursday afternoon, between meetings, with no warning. The S&P jumped four percent in an hour. He cut a third time in November. By the end of the year the equity market had recovered everything it had lost and then some, and by March of 1999 it was making new highs.
But the smart money in late September 1998 did not buy stocks. The smart money bought duration. Two-year Treasury yields had been at 5.5 percent in July. By October 5th they were at 3.85 percent — a hundred and sixty-five basis points lower in ten weeks, the fastest move in the front end since the 1980 recession. The flight to quality made the bond trade first. The Fed cuts confirmed it. The equity rally came later, after the cuts had been delivered and counted.
The pattern is the same every time a leveraged fund cracks open. The bond market moves first. The equity market follows. The Fed shows up at the meeting after that, and the people who bought the front end of the curve when no one wanted it are the ones who collect the check. The ten-year today is at 4.62 percent. The two-year is at 4.18. Watch where they go in the next ten sessions.
When a leveraged hedge fund gates its investors, the answer is not in the equity market. The answer is in the curve. Look at what two-year and ten-year Treasuries did between September 23rd and October 15th, 1998. Then look at what they are doing right now. The first move always comes from the place no one is watching.
◉ TOMORROW’S WATCH
The Sunday futures open in Asia tomorrow night will tell us whether the contagion reaches Tokyo and Singapore the way it did in October 1998, when the yen ripped almost ten percent in two days and took a second wave of carry trades down with it. The quiet thing building today is which prime broker has to stop quoting first.
