Authors: William D. Cohan
——
A
N E-MAIL EXCHANGE
among various members of Goldman’s mortgage securities group on October 19, 2006, illustrated the inevitability of the conclusion Birnbaum and Primer were in the process of reaching about the growing risks in the market. A Goldman mortgage-backed securities sales executive,
Mitchell Resnick, had been running into some resistance from investors increasingly worried about the risks of being long the new crop of
residential mortgage-backed securities (RMBS) then coming to market. “Do we have anything talking about how great the BBB sector of the RMBS is at this point in time?” he wrote Jonathan
Egol, a trader on the structured products desk. “A common response I am hearing on [two deals Goldman was then marketing] is a concern about the housing market and BBB in particular. We need to arm sales with a bit more—do we have anything?”
The response came from
David Rosenblum, another mortgage-trading professional. “So amazing you should ask,” he wrote. “[W]e had this convo for an hour last night—[Alan] [B]razil and [Michael] [M]arschoun and [P]rimer—THIS IS WHAT WE’RE TALKING ABOUT! Can you come to the rescue here?” Five days later, another mortgage trader, Geoffrey Williams, wrote to Egol, in another indication of emerging trouble. “Thinking we need to better leverage syndicate to move open risk from our bespoke trades given that most of them did not go through the initial syndication process,” he wrote. “[G]uessing sales people view the syndicate axe”—being “axed” in a security was Goldman-speak for wanting to sell it, fast—“e-mail we have used in the past as a way to distribute junk that nobody was dumb enough to take first time around.” Egol responded, “LDL,” for “let’s discuss live,” a way for the Goldman traders to avoid writing something in an e-mail that might prove embarrassing later.
The other dynamic in the securitization market that appealed to Birnbaum—and gave him additional confidence that the short side of the bet could be a big winner—stemmed from the nature of creating a collateralized debt obligation in the first place. To do so, Wall Street firms needed the raw material—mortgages and other debt securities—to stuff into a CDO before turning it into a security that could be sold to investors. (Gotta break a few eggs to make an omelet.) For instance, in the case of GSAMP-S2, Goldman had to buy the mortgages from New Century first—since Goldman itself did not originate mortgages. Once Goldman had enough mortgages to bundle them up into a security, the security would get made, and then sold. Often it would take six months or so for enough mortgages to be purchased to be able to make the security. During this period, investment banks, such as Goldman, would keep the mortgages in “warehouses” and collect interest and principal payments from borrowers during the time before the mortgages were made into securities and sold to investors (who then became the beneficiaries of the payments, as long as they were being made). In the meantime, “I have a positive carry trade because these are bonds that I’m earning interest on at a higher rate than where I’m financing them,” Birnbaum said, “and as long as the world doesn’t blow up it’s nice extra P&L”—profit—“for the deal. So the Street loved it, senior managers loved it. So they kept these warehouses out there for long periods of time.”
What Birnbaum realized as he was readying his big bet against
mortgages was that as the mortgages in the warehouses started to lose value, the executives at the Wall Street firms would want to get rid of them as fast as they could, making for a classic supply-and-demand imbalance that would vastly favor the buyer of such an asset over the seller. What was once an asset that provided extra vig in the deal would quickly turn into a liability for firms with big warehouse exposure as they tried to unload the toxic waste. What’s more, many of these warehouses were not even on the balance sheets of the Wall Street firms. It was another lightbulb moment for Birnbaum. “What that means is that you had a huge amount of off-balance-sheet, unquantified risk at the banks,” he said, “and that when, or if, the shit hits the fan and the liabilities of the CDOs—meaning the bonds that the CDO was going to be created from and then sold to investors—you wouldn’t be able to sell those bonds anymore if the market blew up. Then the banks would be hung with all this extra stuff in their warehouses, which wasn’t being quantified anyway and they would wake up and be like, ‘Oh shit, what are we going to do with these warehouses?’ ” That’s when Birnbaum figured the selling pressure would be at its most intense. He figured these banks would conclude, “We’re going to have to liquidate these warehouses, and then they’re all going to be going through the door—a small door at that point—at the same time.”
Compounding the problem would be the fact that buyers would be few and far between, exacerbating the selling pressure. “You would have a massively unraveling event, and a lot of negative momentum,” Birnbaum said. It would be like the final scene in the 1983 movie
Trading Places,
when
Dan Aykroyd and
Eddie Murphy turn the tables on their nemeses the Duke brothers, forcing them into liquidation. The way Birnbaum figured it, Goldman would be plenty happy to buy when others were selling, since by going short it would already have sold the security at a higher price. By buying it back at the lower price, the difference—sell high, buy low—would be pure profit. On Wall Street, this was known as “covering your short” and was a way to make a killing. “That last point was something which was very powerful, because it’s great to say that you’re going to put on a directional trade or any kind of trade and then you’re going to be right, but if you can’t monetize that trade, then there’s no exit strategy,” he said. “Not only was this a way that the market would go down, but it was also the exit strategy because as the CDO managers are liquidating their warehouses, on the way down, you’re sitting there going, ‘Do I want to buy any debt from them or not?’ But at least you have a supply that you can use to cover your shorts. So this was a very important realization and part of the strategy.”
——
B
UT
B
IRNBAUM AND
his colleagues couldn’t just start placing the bet the mortgage market would collapse until they had begun to reduce their previous bets that it would remain robust. “We had pockets of our business that were long the market,” he said. “You can’t get short until you’re at zero. So, the first order of business was taking the pockets of our business that were long the market and getting them closer to home, meaning getting them closer to flat.”
At the same time, the risk-management apparatchiks at the firm were closely monitoring what Birnbaum and company were doing. “Whether you’re long or short, generally the message from risk management is ‘Take less risk,’ ” he said. “They would love you to make infinite profits on zero risk.” One of those apparatchiks—unnamed, of course—defended the practice of working hard to minimize risk as being in the best interests of Goldman Sachs and as part of what makes the firm successful. “I call it both seamless horizontal and vertical communication,” he said, “and not just on the business side, but on the control side as well. We have a lot of checks and balances in place in both the business side and the control side and information sharing. This commitment at Goldman to a rigorous daily mark-to-market, constantly kind of reassessing and checking and really listening to the market. Contrary to popular belief, we’re certainly not smarter than the market. We’re a participant in the market. But having the commitment to daily mark-to-market and having the communication both laterally and up and down makes a huge difference. Having been at
Deutsche Bank and
Morgan Stanley, I can just say here that the way that information flows to people that need that information and the collegial atmosphere in terms of the sharing of that information—you know, businesses aren’t siloed, risk isn’t siloed here—and that’s a big difference from what I’ve seen at other places.”
On October 26, 2006, in a move to reduce the firm’s exposure to the mortgage market, Michael “Swenny” Swenson decided to sell $1 billion worth of the ABX index and to buy another $1 billion of protection, using credit-default swaps on BBB-mortgage securities. “This is estimated to reduce the scenario risk by approx. $90mm,” according to an e-mail circulated in the group. This lowered exposure would no doubt be welcomed news to
David Viniar when Birnbaum and his team met with him to analyze Birnbaum’s risk positions. The message from Viniar throughout 2006 was always the same: take less risk, and make more money doing it.
——
O
NE OF THE
most important responsibilities Viniar had as Goldman’s CFO was to monitor closely the amount of risk the firm was taking on any given day. He looked at the income statements—known as “P&L”s, for “profit and loss”—of Goldman’s roughly forty-five business lines every day. He also made a particular habit of speaking directly to the firm’s risk managers and its head traders to gauge the extent of the dollars they had on the line at any given moment. Not only did Viniar make this part of his daily routine, but so did both
Lloyd Blankfein and
Gary Cohn, Goldman’s president. “
I can give you one hundred different risk-management rules,” Viniar said, “but that’s a good early warning sign—just ‘How are they doing?’ ” These open lines of communication from the bottom of the firm (the traders) to the top of the firm (its CEO, president, and CFO) made Goldman pretty unique on Wall Street.
Like the rest of Wall Street in December 2006, Goldman Sachs was mostly “long” mortgages, meaning that the firm’s traders thought their value would increase over time and so had invested some of the firm’s more than $70 billion in capital in that idea. “
The world was good, right?” Viniar said. “You are long everything because everything went up in value. We’re not very long, but we’re long.… The other thing we do is mark all of our positions to market every day. We’re really diligent about that.”
At one point during mid-December 2006, Viniar began to observe a trading anomaly. “I get the P&L every day, and for something like ten days in a row the mortgage desk lost money,” he explained. “They didn’t lose a lot, but they lost money ten days in a row.” Not a lot of money by Goldman’s standards—somewhere in the $5 million to $30 million range—but a pattern of daily losses that was troubling to Viniar.
Viniar had also been hearing on a regular basis from Daniel Sparks, the Goldman partner who headed up the four-hundred-person mortgage-trading department. Sparks had just been named head of the group in December 2006 and had a seat on the firm’s powerful—and powerfully important—risk-management committee, which met weekly to assess and discuss the firm’s financial risks. He often spoke directly to Viniar and to Cohn and Blankfein about what was happening in the mortgage department. “Subprime market getting hit hard,” he wrote to
Thomas Montag,
William McMahon, and
Richard Ruzika, three of his senior colleagues in the fixed-income division at Goldman, on December 5. “[H]edge funds hitting street.… At this point we are down $20mm today.” Then, in a plug for Birnbaum’s emerging strategy to get shorter more quickly, he added, “Structured exits are the way to reduce risk. Our
prior structured trade closes today. We are focusing on ways to do it again much faster.”
Sometimes these discussions were very frank and direct about the risks the firm was taking in this area. “
One of my jobs at the time was to make sure Gary [Cohn] and David and Lloyd knew what was going on,” Sparks said. “They don’t like surprises, so they need to know in real time if there’s good or bad things happening. They were the kind of guys that I think were good enough managers so if you tell them that, then they can deal with it.” Sparks was becoming increasingly nervous about Goldman’s mortgage portfolio. He was worried that loans Goldman had made to mortgage originators, like New Century (which owed Goldman millions), were not being repaid on a timely basis. He was also worried about the warehoused mortgages Goldman had bought that had not yet been turned into securities and that borrowers were increasingly defaulting on. The mortgage originators had an obligation to buy back from Goldman the mortgages that were not performing; this was not happening either. “
We were seeing signs in the department level that were concerning,” Sparks said. “The firm had exposure.… David asked me to put together basically a review of all of the different risks that we had in the department and there were a lot of risks.” Birnbaum credited Sparks with realizing that the mortgage originators—particularly New Century—were not paying their debts as they became due and decided to cut off their access to credit from Goldman. “
That was kind of a first sign,” Sparks said, “but it also potentially meant that a lot of other bad things could be happening, too.”
After hearing repeatedly from Sparks about his growing concerns, Viniar called a meeting, for December 14, of the five people running the firm’s FICC group—fixed-income, currencies, and commodities—and the various controllers, auditors, and risk managers who work with them in those divisions. In sum, about twenty people had collected around Viniar in his thirtieth-floor conference room at 85 Broad Street for one of the most momentous meetings in Wall Street history.
The mortgage traders came to the meeting with a two-inch-thick report detailing all the firm’s mortgage-related trading and credit positions. Viniar said the firm had not made big bets one way or the other, but rather had a series of bets that tilted toward prices going up or down. At that moment, Goldman’s bias was toward betting the value of the trades and the underlying mortgages would increase. But even so, there were disputes about the value of the mortgage securities with some of Goldman’s trading partners.
The meeting dragged on for close to three hours. Each position was
reviewed and then reviewed again. The firm had lost money ten days in a row because, Viniar said, it was betting mortgage-based securities would rise, when “the market was going down.” They talked about the big bet that John Paulson—the hedge-fund manager—had made that the mortgage market would collapse. As the meeting was winding up, Viniar concluded by saying, “It feels like it’s going to get worse before it gets better.” Looking back on the meeting, he commented, “Nobody there knew how bad it was going to get. We didn’t have a clue that things were going to go crashing down.” There was as close to complete consensus in the room about Viniar’s conclusion as was possible. “A pretty good consensus developed that we needed to reduce our risk,” said someone who was at the meeting.