The Big Short Chapter 7: The Great Treasure Hunt for Bad Mortgage Bonds
Chapter 7 of The Big Short is called “The Great Treasure Hunt,” and I think it is the most frustrating chapter in the whole book. Not because it is boring. Because it shows you that every institution that was supposed to protect the system - the rating agencies, the regulators, the big banks - was either clueless, corrupt, or both.
By the time we get here, our characters already know the trade. They already shorted the subprime bonds. Now the question becomes: how deep does the rot actually go? And the answer, as they discover one conversation at a time, is basically all the way down.
Cornwall Capital and the Disappearing Parachutes
We pick up with Charlie Ledley and Ben Hockett from Cornwall Capital, the garage-band hedge fund. They just came back from the Las Vegas conference in late January 2007 completely convinced that the financial system has lost its mind. Charlie tells his own mother he thinks they might be facing “the end of democratic capitalism.” Her response? She suggests he go on lithium.
Which, honestly, is kind of how everyone reacted to these guys.
Cornwall scrambles to buy as much insurance on subprime bonds as possible. They get an ISDA agreement with Morgan Stanley pushed through in ten days, which normally takes months. Charlie stays up nights figuring out which CDOs to bet against. On February 16, 2007, after fighting over the price, they pay 150 basis points to buy $10 million in credit default swaps on a CDO called “Gulfstream.” Whatever that was.
Then something wild happens. Five days later, the market starts trading an index of CDOs called the TABX. For the first time ever, you can see the price of these things on a screen. And the double-A-rated tranche that Cornwall bet against? It closed at 49.25. Lost more than half its value on the first day.
Think about this. Wall Street was selling these same bonds at 100 - full price, everything is fine. And simultaneously, the index made of the same bonds was trading at 49 cents on the dollar. Two hands of the same firm doing completely opposite things.
Charlie calls Morgan Stanley the next morning to buy more. “I’m really, really sorry,” the woman tells him, “but we’re not doing any more of this. The firm’s changed its mind.” Overnight. Her boss said something vague about a risk management decision at “the very highest levels of Morgan Stanley.”
Nobody would tell them what happened. But Cornwall found Wachovia, who was still willing to sell cheap insurance but too nervous to deal with Cornwall directly. So Bear Stearns sat in the middle as a broker. A $45 million trade agreed in February didn’t go through until May. Ben Hockett described it perfectly: “It was like we were in a plane at thirty thousand feet, which had stalled, and Wachovia still had a few parachutes for sale.”
After that, the market shut down. Cornwall now owned $205 million in credit default swaps on a portfolio of less than $30 million. They were disturbed mainly that they didn’t own more.
The Rotten Oranges and the Fresh Juice
Here is the thing that should make your blood boil. Between February and June 2007, even as the subprime mortgage bonds were clearly falling apart, Merrill Lynch and Citigroup created and sold $50 billion in new CDOs. Fifty billion.
The underlying bonds had lost 30 percent of their value. If the oranges are rotten, the orange juice should be rotten too. But the CDO machine kept running. Cornwall was “totally baffled.” As Charlie put it, “Everyone and everything just goes back to normal, even though it obviously wasn’t normal.”
Bear Stearns published research saying the declines were just “market sentiment.” They held a CDO conference where a presentation titled “How to Short a CDO” was quietly removed from the agenda. Moody’s and S&P said they were reconsidering their ratings models for new bonds, but when Cornwall’s lawyer asked them to reconsider the two trillion dollars of bonds already rated badly, the answer was “Hmmmmmm…no.”
By late March 2007, Charlie and his partners were pretty sure of one of two things. “Either the game was totally rigged, or we had gone totally fucking crazy.”
They tried the press. The New York Times and Wall Street Journal reporters they knew weren’t interested. They went to the SEC enforcement division. The SEC listened politely but clearly didn’t understand what a CDO was. “It was almost like a therapy session,” said Jamie. The SEC never followed up.
Eisman Goes Looking for Hidden Garbage
Now we switch to Steve Eisman, who by this point is short about $600 million in subprime securities and wants to short more. His risk management people at Morgan Stanley keep calling his trader Danny Moses, asking him to reduce the position. Danny would go to Steve, Steve would say “Just tell them to fuck off,” and Danny would relay the message.
His team even started hiding information from him. Lippmann at Deutsche Bank was sending negative housing data, and Danny and Vinny were worried Eisman would shout “Do a trillion!”
In March 2007, Bernanke said the subprime problem “seems likely to be contained.” Eisman was disgusted. “Credit quality always gets better in March and April. People get their tax refunds.” He thought the market was moronic for taking this seasonal blip as a real signal.
Danny Moses turned off CNBC. “If something negative happened, they’d spin it positive. It alters your mind.”
The Rating Agencies Had No Data
Here is where it gets really bad. Eisman starts investigating what the rating agencies actually know. He discovers something that I still find hard to believe even years later.
Floating-rate subprime mortgages were structured so the borrower paid a teaser rate of 8 percent for two years, then the rate jumped to 12 percent. These were people “one broken refrigerator away from default.” Obviously, they were more likely to default when the rate jumped. But the rating agencies assumed borrowers were just as likely to pay at 12 percent as at 8 percent. So bonds backed by floating-rate mortgages got higher ratings. Which is why floating-rate loans went from 40 to 80 percent of subprime mortgages. The system was built around the model, not reality.
Eisman took the subway to S&P’s office to meet with an analyst named Ernestine Warner from the surveillance department. The surveillance department was supposed to monitor bonds and downgrade them when the underlying loans went bad. The loans were going bad. The bonds were not being downgraded. Eisman wanted to know why.
What he found was insane. The rating agencies were working with the same rough data available to traders like Eisman. They didn’t have the loan-level detail. They couldn’t see which specific loans were insured, which were likely to default. When Vinny asked Warner why, she said, “The issuers won’t give it to us.”
Vinny lost it. “You need to demand to get it! You’re the grown-up. You’re the cop! Tell them to fucking give it to you!!!”
But S&P was afraid. If they demanded better data from Wall Street, Wall Street would just take its business to Moody’s. The cops were afraid of the criminals.
Eisman also visited Moody’s CEO, Ray McDaniel, who told them with a straight face, “I truly believe that our ratings will prove accurate.” As they left, Vinny said politely, “With all due respect, sir, you’re delusional.” Vincent Daniel, from Queens, telling the CEO of Moody’s - 20 percent owned by Warren Buffett - that he’s delusional. I love this moment.
The Treasure Hunt Begins
By summer 2007, Eisman’s positions were moving in his favor by millions of dollars a day. But he kept asking the most important question: “I know I’m making money. So who is losing money?”
He started looking at the big Wall Street banks themselves. His original thesis was that when the securitization machine died, their revenues would dry up. But then something worse occurred to him. What if the banks weren’t just selling this garbage? What if they were keeping some of it?
HSBC announced surprise losses. Merrill Lynch admitted subprime losses in its quarterly results. These were signals that the biggest firms on Wall Street were sitting on piles of the very toxic waste they had been selling.
Eisman started meeting with Wall Street CEOs, asking basic questions about their balance sheets. “They didn’t know,” he said. “They didn’t know their own balance sheets.” He met Bank of America’s CEO Ken Lewis. “I had an epiphany. ‘Oh my God, he’s dumb!’ The guy running one of the biggest banks in the world is dumb!”
They shorted Bank of America, UBS, Citigroup, Lehman Brothers. Would have shorted Morgan Stanley too, but they were owned by Morgan Stanley.
At a meeting with Merrill Lynch’s biggest shareholders, their CFO explained everything was under control, the models said so. Eisman interrupted: “Well, your models are wrong!” Then he started conspicuously stacking his papers as if to leave. The room went silent.
A Newsletter Editor Who Actually Read the Documents
Meanwhile, Jim Grant of Grant’s Interest Rate Observer sent his assistant Dan Gertner, a chemical engineer with an MBA, to read the CDO documents. Gertner sweated over them and came back saying he couldn’t figure them out. Grant’s response: “I think we have our story.”
If a chemical engineer with an MBA couldn’t understand what was inside a CDO, neither could the investors buying them. Grant published pieces saying the rating agencies had abandoned their posts. S&P summoned him to their office and told him he “just didn’t get it.”
When Eisman read Grant’s work, he said it was like “owning a gold mine.” Independent confirmation from someone who actually looked at the data.
What This Chapter Is Really About
This chapter is about a search for information. Our characters are looking at a system where everyone - the banks, the rating agencies, the regulators, the media - has stopped asking basic questions. Nobody wants to know what is actually inside these bonds. Nobody wants to see what happens if housing prices fall. Nobody wants to demand the data.
The people who did ask - Cornwall Capital, Eisman’s team, Jim Grant - were treated like cranks. Their own mothers suggested medication. The SEC couldn’t follow the explanation. Reporters didn’t want the story.
In my 20 plus years in technology, I have seen this pattern many times. Not at this scale. But the dynamic is the same. When a system is making money for everyone, nobody wants to ask uncomfortable questions. The incentives all point toward looking the other way.
The saddest part? The information was there. The rating agencies could have demanded loan-level data. The SEC could have investigated. The banks could have looked at their own balance sheets. Nobody wanted to look. And when a few people forced them to look, they were told they “just didn’t get it.”
They got it. The rest of the world was about to find out.
Previous: Chapter 6: Spider-Man at The Venetian