Authors: William D. Cohan
——
T
HE PROBLEMS AT
New Century and Fremont quickly began to ripple through the market. On March 12, as previously instructed, the ABACUS deal team, including Egol and Tourre, presented ABACUS to Goldman’s “Mortgage Capital Committee” to get its approval. According to the memo about the deal, Goldman stood to make between $15 million and $20 million for acting as an intermediary between Paulson and ACA. There appeared to be little discussion of the reputational risk the firm might suffer as a result of the deal, which had been the reason Tourre had been instructed to get the committee’s approval even though Goldman did not appear, at first, to be committing any capital. Indeed, according to Egol and Tourre, ABACUS had it all. “This transaction is a new and innovative transaction for Goldman Sachs and the CDO Market,” they wrote. After noting the deal’s highly technical “firsts,” they continued, “This transaction addresses the objectives of multiple clients of the firm: it helps ACA increase [its] assets under management and [its] fee income; it enables Paulson to execute a macro hedge on the RMBS market; it offers to CDO investors an attractive product relative to other structured credit products available in the market. Our ability to structure and execute complicated transactions to meet multiple clients needs and objectives is key for our franchise.” The committee approved the deal.
That afternoon,
Jörg Zimmerman, a vice president at IKB Credit Asset Management, a big German bank in Düsseldorf that was taking the long side of the ABACUS deal, wrote to Michael Nartey, the Goldman banker in London, with copies to Tourre and Egol, that IKB wanted to remove both the Fremont and New Century bonds from the reference list for the ABACUS deal, no doubt because of the two companies’ ongoing financial difficulties. Zimmerman wrote that he wanted to go back to IKB’s “advisory comitee [
sic
]” and “would need consent on” removing these securities from the ABACUS deal. This was not such great news. “Paulson will likely not agree to this unless we tell them nobody will buy these bonds if we don’t make that change,” Tourre wrote to Egol, who wrote back wanting to know what “we say to Joerg [
sic
]?” “As discussed with Nartey,” Tourre replied, “we are taking his feedback into account and once we have gotten more feedback from accounts across the cap
structure we will decide what the best cours[e] of action is.” Tourre’s head fake was typical of bankers looking to make it seem there was competition for a deal when clearly there was not. Indeed, IKB may have been one of the few investors the world over willing to take the long side of such a trade with so many red flags emerging about the problems in the mortgage market. (There is no additional documentation about whether Goldman agreed to take out the New Century and Fremont mortgages, but the final ABACUS deal did include mortgages serviced by both companies in the reference portfolio; Zimmerman did not respond to a request for an interview.)
New Century’s problems were also giving pause to Rabobank, a big Dutch bank, which was considering investing in
Anderson Mezzanine Funding, another Goldman-architected $305 million CDO also coming to market in March. Unlike the ABACUS deal, in the Anderson deal, Goldman was underwriting the equity portion and expected to keep half of it as a principal investment. This prompted
Olivia Ha, a 1998 Harvard graduate and Goldman vice president, to e-mail the Anderson deal team at Goldman with a question about how it got comfortable with the New Century collateral since Ha’s client at Rabobank,
Wendy Rosenfeld, had expressed her concerns about it. “[H]ow did you get comfortable with all the [N]ew [C]entury collateral in particular the [N]ew [C]entury serviced deals[,] considering you are holding the equity and their servicing may not be around[?] [I]s that concerning to you at all?” Rosenfeld needed “more comfort” because she was “getting credit resistance on the [N]ew [C]entury concentration.” Eventually, several members of the Anderson deal team at Goldman got on the phone with Wendy to “allay her [N]ew [C]entury concerns.… This will be our opportunity to help arm her with ammo for her credit [committee] who is getting jittery on the [N]ew [C]entury exposure/servicing concentration.”
The call between Goldman and Rosenfeld did not go as Goldman had hoped. She ended up passing. “At this point in time we are not going to be able to participate in Anderson,” she wrote to Ha on March 13. “There are many concerns regarding the percent of NC”—New Century—“originated and serviced collateral.” A few days later, several other potential Anderson investors dropped out, also because of “New [C]entury issues.”
Smith Breeden Associates, a global asset management company, dropped out as well over concerns the deal would get downgraded and because there would not be sufficient cash flow to cover the interest payments.
Scott Wisenbaker, the Goldman banker on the deal, agreed to speak with Smith Breeden to “make sure they understand the deal correctly, but regardless, it looks like they are lower probability to be involved.”
This was not the answer
Peter Ostrem, a more senior Goldman banker and the head of the CDO group, was looking to hear. “Yeah?” he fired back to Wisenbaker. “So—fix the miscommunication so the probability goes up.”
Nerves seemed to be getting frayed throughout Goldman. There was growing pressure to deal with what appeared to be the increasing likelihood of a financial crisis brought on by the problems in the mortgage market. On March 14, the firm sent around an internal economic research report that contained an interesting nugget that caught the attention of the firm’s senior executives and got them even more worried than they already were. “New data from the
Mortgage Bankers Association show that mortgage credit quality problems go well beyond the subprime sector,” according to the report. “This can be seen from the fact that delinquencies on prime adjustable-rate mortgages are rising quickly—much more quickly, in fact, than those of subprime fixed-rate loans.” When Winkelried read this he grew concerned. “[D]elinquencies in PRIME adjustables with teaser rates growing fast,” he wrote to Blankfein, Cohn, and Viniar, plus Ruzika and Sparks, the two guys running the mortgage group. “[I] think this may be a big problem and a lot worse than currently thought. [I] think lending standards are highly variable among originators and [W]all [S]treet focus (servicers and dealers) on quality control has been lost for a while. [D]an, are we doing things to prepare for bleed into prime space?”
Sparks responded quickly. “Trying to be smaller and buying puts on companies with exposure to overall mortgage market,” he wrote to the same august group. “We are also short a bunch of sub-prime AAA index.” Goldman’s schizophrenic behavior continued, though. The same day that Sparks was writing to Winkelried, Cohn, and Blankfein about being short the ABX index and buying puts on companies with exposure to the mortgage market, he also sent Montag an e-mail titled “Cactus Delivers,” about
Mehra Cactus Raazi, a Goldman bond salesman and former
Rolling Stone
ad salesman, urging Montag to congratulate Raazi on selling off a $1.2 billion short position the firm had in A-rated mortgage securities to
Stanfield Capital Partners, a New York–based CDO manager. “He did a great job filling our ax,” Sparks wrote to Montag, who then sent the whole e-mail chain on to Blankfein with a note: “Covered another [$]1.2 billion in shorts in mortgages[,] almost flat”—a reference to being neither too short nor too long the mortgage market—“now need to reduce risk.” On March 20, after Blankfein received the daily firmwide “net revenues” estimate that showed the firm had generated $111 million in revenues that day—and $37 million in pretax earnings—but had lost
$21.4 million in mortgages, he sent an e-mail to Cohn. “Anything noteworthy about the losses in mortgages?” he wanted to know. Cohn replied, “No[.] [M]arket rallied a bit[.] [S]till short.”
The next day, this little executive-level colloquy trickled downstream.
“Did Josh get out of index trade?” Montag asked Sparks and Bill McMahon, a reference to Birnbaum’s short position on the ABX that had started to—briefly—move against him. “I had him liquidate S&P’s”—a bet made on the S&P 500 stock index—“and cut equity put position in half yesterday.” He then explained what the group was focused on, strategically, which no doubt pleased the VAR police but was equally disheartening to Birnbaum. “Overall as a business, we are selling our longs and covering our shorts,” Sparks wrote, “which is what this quarter is really about, as well as protecting ourselves on counterparty risk, planning for the new resi[dential mortgage] world, and trying to be opportunistic. We have shorts that we need to provide overall protection in case we get further move downs—and those shorts have been hurting us.”
Sparks’s logic may have been flawless, but Montag had little patience, it seemed, for a bet that might pay off down the road but that was moving against the firm in the short term, especially after Blankfein had made his inquiries. “Liquidity is better,” Montag conceded, “but actual performance can be much worse obviously.” Then he took a swipe at Birnbaum: “Unfortunately[,] the trader [J]osh has not demonstrated a track record of controlling his position.… Instead of these lousy hedges he should just be selling his position.” Sparks tried to defend Birnbaum. “He has had a very good run in this activity,” he wrote to Montag and promised that before long Birnbaum would “lay out [a] plan” for how to proceed and get Montag to sign off on it.
On March 26, Goldman management gave the Goldman board of directors a presentation on the subprime mortgage market. The twenty-four-page document contained a page titled “The Subprime Meltdown,” which traced the collapse of New Century’s stock and the financial carnage among the subprime mortgage originators, including the fifteen companies that had already been liquidated or filed for bankruptcy. The presentation also described Goldman’s dual, schizophrenic role in the market: one as a buyer, packager, and seller of mortgages and mortgage-related securities to investors for a fee—Goldman “exits loan purchases by structuring and underwriting securitization and distributing securities back by mortgage loans on a principal basis and for clients,” the presentation
explained—the other as a trader of mortgages and derivatives related to mortgages “to hedge our long credit exposure” in a bet that the mortgage market would collapse. It was quite a pas de deux.
Goldman’s management also created a timeline of the firm’s reactions to worsening market conditions in the
subprime mortgage sector. For instance, in the second half of 2006 and the first quarter of 2007, Goldman “reduces CDO [origination] activity” and “residual assets marked down to reflect market deterioration.” Then, the board was told, “GS reverses long market position through purchase of single name CDS”—credit-default swaps—“and reductions of ABX.” The gross revenues of Goldman’s mortgage business reflected the changing dynamics as well. In 2005, the firm made $885 million in revenues on the mortgage desk, mostly from the origination of residential and commercial real-estate securities. In 2006, Goldman underwrote $29.3 billion of subprime mortgage securities, a ranking of sixth overall, and underwrote close to $16 billion in collateralized debt obligations, ranking fifth. That year, the mortgage-related revenues increased 16 percent with the origination business staying essentially flat, but with Birnbaum’s group generating $401 million in revenue, up 64 percent from the $245 million generated the year before. That revenue generation accelerated in the first four and a half months of fiscal 2007, where Birnbaum’s group had produced $201 million in revenue, already half of what had been generated for all of 2006. Meanwhile, the residential mortgage security origination business had fallen off the cliff by the first part of 2007, with a $19 million loss in gross revenue. The Goldman board also was shown that the firm was long some $12.9 billion in various mortgage securities, which was offset by a $7.2 billion short bet against the ABX and another $5.5 billion negative bet against mortgages, obtained through the purchase of credit-default swaps. Goldman’s net exposure as of March 15, 2007, was some $200 million on the long side, or virtually flat. In March 2007, after Goldman’s first-quarter performance was released, Viniar said, “Subprime is under stress, it appears to have been overheated. It’s pretty clear there will be a shakeout. It will be a reasonably sized, but smaller market than it has been over the last several months.” As to the firm’s ongoing role in the mortgage market, he said, “When we extend credit we tend to have security and other terms that will protect us. We do what we can to mitigate our losses, we do what we can to protect ourselves.”
Incredibly, many others—among them
Ben Bernanke, the Fed chairman, and Henry Paulson, the treasury secretary—were missing the problems that Goldman Sachs and
John Paulson were seeing in the
mortgage market. “The impact on the broader economy and financial markets of the problems in the subprime market seems likely to be contained,” Bernanke testified before the
Joint Economic Committee of Congress on March 28. That same day, Paulson told a House Appropriations subcommittee: “From the standpoint of the overall economy, my bottom line is we’re watching it closely but it appears to be contained.”
Bear Stearns was also projecting a very different outlook on the opportunities in the mortgage market than was Goldman. In a March 29 “Investor Day” presentation,
Jeffrey Mayer and
Thomas Marano, two of the most senior executives in Bear’s fixed-income group, were trumpeting the fact that the firm’s “leading mortgage franchise continues to grow.” The hits just kept on coming: net revenues had doubled since 2002 to $4.2 billion in 2006; Bear was the “top ranked” underwriter of mortgage-backed securities and asset-backed securities; the firm had expanded its mortgage origination capabilities by purchasing
Encore Credit Corporation—a “sub-prime wholesale originator”—to complement
Bear Stearns Residential Mortgage Corporation and
EMC Mortgage Corporation; and Bear ranked fifth in the underwriting of CDOs, with a volume of $23 billion in 2006, with “volume nearly doubling from last year.” The men also boasted of being “well-positioned to handle disruption in the sub-prime market.”