Authors: William D. Cohan
Birnbaum was not pleased. He knew his bets would be proved right—and be worth many billions of dollars—and yet the corporate drones were clipping his wings. “There was a certain amount of tension regarding just how short should our group be,” he said. “Because we’re
the only ones who were taking short risk at the firm level. Just how short should we be? If you were evaluating us as a stand-alone, then we were pretty darn short. If you were evaluating us in the context of some of these other positions, there were some of us who felt that we weren’t short enough, and that the quantification of those other desks wasn’t being done enough. And the quantification of our desk was done ad infinitum.” If he had his own hedge fund, like
John Paulson or
Kyle Bass, his potential payday would be nearly unfathomable. According to the
Wall Street Journal,
he “briefly lost his cool and slammed down a phone receiver … when a more senior bond trader insisted on unwinding some of his trades to cut risk.” He referred to the audit gnomes from Goldman’s controller’s office as the “VAR police” because they were constantly pestering him about the level of risk he was taking. “The sexiest job within accounting at Goldman Sachs was to be focused on the structured products group,” he said. “Because the P&L swings were huge. For the most part they were positive swings.… It was the Super Bowl if you were in accounting.”
Birnbaum saw tremendous irony in the way the “VAR police” reacted to the money Goldman was making from his short bets. The mortgage market would move down, Birnbaum’s bets would be more valuable, and Goldman would make more money. But that also meant the market had become more volatile, which is one of the significant variables in the VAR computer model. Birnbaum’s desk would make $1 billion in a given month in 2007 and still own the same positions it owned the previous month but then be told the risk had increased. “You had the same position deemed to be twice as risky after you made a billion dollars in one month,” he said. “And then someone will come knocking on your door and say, ‘Oh, you know, you’re taking massive amounts of risk. You need to cut your risk profile.’ ”
Josh Birnbaum did not appreciate being singled out. He urged the risk managers to instead take a more unified approach and look at the VAR across the spectrum of Goldman’s bets, long and short, not just his moneymaking shorts. Alternatively, if his ability to make the short bet would be limited from time to time and compared with what was happening with the long bets, let him manage the long bets and then hold him accountable for those, too. “It’d be one thing if they said, ‘We’re holistically looking at this. We realize that you’re massively short. These guys are long. And we’re where we want to be.’ Instead it was, ‘Don’t worry about those other businesses. That’s our problem. You’re massively short. Your VAR is on fire. Cut your risk,’ ” said one person with knowledge of this dispute.
So that is what Birnbaum and his team did, much to their chagrin. On February 25—a Sunday—Sparks e-mailed Montag with a progress report on the trading desk’s efforts to reduce their risk. He informed Montag that the desk had covered $2.2 billion in short positions obtained by buying credit-default swaps but also had sold short $400 million worth of the BBB-ABX index. “Desk is net short,” he wrote, “but less than before. Shorts are in senior tranches of indexes sold and in single names. Plan is to continue to trade from the short side, cover more single names and sell BBB-index outright.” Sparks also let Montag know that some $530 million worth of mortgages and mortgage securities that Goldman had been storing in inventory to create new CDOs had been liquidated, and that another $820 million warehouse had begun liquidating. He wrote that after this wave of liquidations, Goldman still had in its pipeline to be sold $2 billion worth of “high-grade deals” and another $2 billion of BB-rated, “CDOs squared,” or CDOs composed of other CDOs, a real dog’s breakfast of risk. “How big and how dangerous” were the CDO squareds? Montag wanted to know. “Roughly [$]2 bb, and they are deals to worry about,” Sparks wrote. One Goldman CDO deal—set to be priced on February 26—was instead liquidated per Sparks’s order. “Thought we’d announce deal tomorrow,” one Goldman banker wrote to a colleague on February 25, “but if we’re just going to liquidate that doesn’t really make sense.”
On February 27, Sparks again turned up the heat on Birnbaum, Swenson, and company to reduce the desk’s risks. He explained that his business’s VAR was up due to volatility in the market but that the “[b]usiness [is] working to reduce exposures” and that “a lot of shorts already covered,” including $4 billion in shorts on single-name mortgage-backed securities. “[The] [b]usiness [is] continuing to clear out loans,” he informed his colleagues.
On the other hand, the origination business was still booming, including the pricing of the firm’s largest-ever commercial mortgage-backed security deal that week. “The deal was oversubscribed,” Sparks said. He also mentioned an $11 billion commercial real-estate loan. “The deal was very well subscribed,” he said. Sparks did not mention to his colleagues that the $2 billion ABACUS deal was still moving forward, but that same day an internal memo was circulated outlining the deal’s “marketing points.” Among them, “Goldman’s market-leading ABACUS program currently has $5.1 billion in outstanding [bonds] with strong secondary trading desk support.” The memo said the ABACUS deal would be priced and sold the week of March 5.
On Saturday, March 3, Sparks wrote himself another e-mail summarizing
“[t]hings we need to do,” including focusing on Goldman’s loan exposures to mortgage originators having trouble and speaking to “sales and clients about our deals.” Many months into “the big short,” this still had not been done, despite Whitehead’s famous first principle. He wondered if the “junior people [are] OK” given all the turmoil in the markets. For his traders, he wanted to make sure they knew not to “add risk,” to “trade everything from short to flat,” to “get out of everything,” and to “discuss liquidity of hedges.” One consequence of Sparks’s ruminations was to consider seriously terminating the ABACUS deal, then on the verge of being priced and sold. Per Sparks’s instruction, on Sunday,
Jonathan Egol e-mailed much of the mortgage-trading group: “Given risk priorities,
subprime news and market conditions, we need to discuss sidelining this deal in favor of prioritizing [another deal] in the short term.” To those people getting ready to syndicate ABACUS, he wrote, “[L]et’s discuss the right way to communicate this internally and externally” and to Tourre, whom he called “Fabs,” he wrote, “[L]et’s focus Paulson on trades we can print now that fit.” Tourre was none too pleased that his hard work was about to be rendered worthless. “Maybe we could have discussed live first before sending this out,” he wrote Egol, with thinly veiled anger. But Egol was not sympathetic. “This is per [S]parks’ instructions,” he shot back.
Tourre quickly became despondent at this news, as became clear in an e-mail correspondence he had with Serres on March 7. “[T]he summary of the US subprime market business situation is that it is not too brilliant …,” he wrote to her. “According to Sparks, that business is totally dead, and the poor little subprime borrowers will not last so long!!! All this is giving me ideas for my medium term future, insomuch as I do not intend to wait for the complete explosion of the industry and the beginning of distressed trading, I think there might be more interesting things to do in Europe.” Tourre told his girlfriend that he had been speaking with Michael Nartey, a managing director in London, “who naturally confirmed that ‘he would love if I were in London, which would greatly facilitate communication with New York and would push the European sales force to concentrate on the risks of structured finance.’ ” He told her he was heading to London in April to “get a better sense for the opportunity but I am getting more and more convinced.” He signed off with his usual affection and added, “I don’t want to give you false hopes but I have a good feeling” about the new job happening.
At this prospect—that Tourre might soon move back to London—Serres was nearly euphoric. “Oh sweetheart, by just implying that you have a good feeling about coming over here, I’M JUST THE HAPPIEST
WOMAN ON EARTH!!!!” She explained how she had had an intense physical workout that morning. “But reading your e-mail, knowing that I can hope to in a day not too far off, wake up in your arms every morning, see the love of the whole world in your eyes and reciprocate it hundredfold, every day … It’s a last generation dose of amphetamines! I ADORE YOU FAB. Can’t wait to whisper sweet words in your ears in a few hours.” To which he responded, “And right now, I’d love nothing more than just curl in your arms, feel the warmth of your skin and just stay there smiling for hours (with occasional—frequent tender kisses) … Wake up slowly my love.” Tourre had more good news, too: the ABACUS deal had stayed on track.
T
here was no question that
Goldman’s top executives were monitoring closely what Sparks was doing to reduce the firm’s mortgage exposure. In a March 5 e-mail circulated to the firm’s top brass—including Blankfein, Cohn, Winkelried, Viniar, John Rogers, and
J. Michael Evans, a vice chairman of the firm who also ran Goldman’s business in Asia—various world market indices had shown some improvement, a fact noted by Evans when he sent the news around. “Feels better,” Cohn replied to Evans. “But anything with a + would feel better.” “Agreed,” Evans wrote, “and the bigger the plus the better.” But, Cohn pointed out, that might not necessarily be true. “A big plus would hurt the Mortgage business but [a trader] thinks he has a big trade lined up for the morning to get us out of a bunch of our short risk.”
Early in the morning on March 8—12:50 a.m.—Sparks sent the senior executives of Goldman, aside from Blankfein, an e-mail summarizing the firm’s mortgage risk. He explained that the firm still had significant exposure on the long side of the mortgage market, including more than $4 billion in CDOs on the books that the firm was trying to get out of—“We have various risk-sharing arrangements, but deal unwinds are very painful,” he wrote—some $4.3 billion of Alt-A home mortgages waiting to be turned into securities, plus $1.3 billion in
subprime mortgages and $700 million in second mortgages. “This market is also very difficult to execute in,” he explained. There was also another $1.65 billion in other
mortgage-related securities. “If the credit environment significantly worsens, these positions will be hurt by losses, further lack of liquidity and lower prices,” he continued. Then, there was the ongoing process of covering Birnbaum’s short bets. “We have longs against them, but we are still net short,” he wrote. There was $4 billion worth of shorts on “single name subprime” and another $9 billion in shorts betting against the ABX index. The shorts, he wrote, “have provided significant
protection so far, and should be helpful … in very bad times.” But, he added, addressing the concerns of the firm’s senior executives, “there is real risk that in medium-[term] moves, we get hurt in all three parts of the business—the long CDOs, the other mortgage-related securities and on the short positions. Therefore, we are trying to close everything down, but stay on the short side. But it takes time as liquidity is tough. And we will likely do some other things like buying puts on companies with exposures to mortgages.”
This last bit—about betting companies with mortgage exposure would collapse—was an interesting new development and the first concession from Goldman that it was hedging itself and would soon be betting that other companies—even some of its competitors—would fail. At that time, Goldman had purchased $60 million, notionally, of equity put options “on subprime lenders” as “risk mitigant to overall subprime business.” Betting against Goldman’s competitors would follow soon enough.
——
D
ESPITE
G
OLDMAN’S
thirtieth-floor decision to do everything it could as a firm to hedge its billions of dollars of exposure—and hence risk—to the mortgage market in December 2006, the firm kept right on packaging up, underwriting, and selling mortgage-related securities of all stripes and sizes: subprime mortgages,
Alt-A mortgages,
home-equity loans, as well as more complicated CDOs and
synthetic CDOs. This activity continued throughout the first half of 2007 until the collapse of the Bear Stearns hedge funds in the early summer of 2007 made that activity nearly impossible. Goldman continued to generate fees underwriting and selling mortgage-related securities at the same time the firm had made the corporate decision to hedge its bets and “get closer to home.” According to a presentation made to the Goldman board of directors in September 2007, Goldman had underwritten $4.4 billion of subprime mortgage-related securities to date, seventh in the league tables, just ahead of Bear Stearns. In CDOs, Goldman put together twelve deals in 2007, totaling $8.4 billion, fourth overall but light-years behind Merrill Lynch, which underwrote $72.5 billion of CDOs in 2007. This certainly appeared to be a clear conflict of interest—betting against the mortgage market as principal at the same time as the firm continued to underwrite mortgage securities as agent.
For his part, Sparks said there was no “
bright line” demarcating the decisions to “get closer to home,” as Viniar said, and the one to keep packaging up the mortgages Goldman had already bought and selling them as mortgage-backed securities as if nothing had changed. “
I don’t think it was ever that cut-and-dried,” he said. “[T]he firm was willing to
sell the mortgage securities cheaply and there were still a lot of investors who wanted to buy them. That was part of the plan, and I don’t think anybody thought the world was coming to an end. The firm was just trying to cut its risk.” Indeed, at that time, Goldman still had plenty of people at the firm whose sole job was to buy mortgages, package them up, and sell them into the market. “That’s kind of what the business was,” he said, adding that to get rid of the mortgages that had been warehoused, Goldman sold the mortgage securities cheaply to investors who wanted to buy them at the prices being offered, and that Goldman ended up keeping the riskiest tranches that could not be sold. “
We lost a ton of money on those deals,” he said. “Like we lost a
lot
. A lot of guys [at other firms] decided not to sell because they didn’t want to take the losses. We said, ‘Okay, we’re going to sell and we’re going to take the loss.’ ” By taking these losses, Goldman’s mortgage desk lost money in the second quarter of 2007—“the only mortgage department on the Street that lost money” in the second quarter, he said—but positioned itself well for the coming calamity. When the news about the losses in the mortgage department in the second quarter was shared at the quarterly internal “town hall” meeting, one trader in the group said he felt like a failure. “I felt like I was the worst trader on the Street, the worst businessman on the Street from a risk-management perspective,” he said. “The reality was we were just doing what we thought was right.”