Money and Power

by William D. Cohan

  “I know, right … model def[initely] does not capture half the risk,” she replied.

  “We should not be rating it,” he answered.

  “We rate every deal,” Mooney replied. “It could be structured by cows and we would rate it.”

  Answered Shah, “[B]ut there’s a lot of risk associated with it. I personally don’t feel comfy signing off as a committee member.”

  In a little-known lecture, in October 2008 at Harvard University, Lew Ranieri said he began to realize “in late 2005” that the markets for mortgage-backed securities had entered the “insanity” phase when homes were no longer being viewed as shelter but rather as a “new form of ATM machine” where home owners were borrowing larger and larger sums against their homes, all the while hoping their value would continue to increase. Ranieri reminded his Harvard audience—at the Graduate School of Design, no less—that not only could housing prices fall occasionally but also that interest rates had a habit of increasing after a period of historic lows. This double whammy is precisely what happened, of course, and it helped to precipitate the crisis. Ranieri, who had once aspired to be an Italian chef but was asthmatic and could not breathe in a smoky kitchen, said of his financial baby that he had “wanted to make a lasagna but ended up with a bouillabaisse.”

  ——

  IN THE MONTHS following Goldman’s March 2006 sale of the GSAMP securities, a U.S. affiliate of Deutsche Bank, the large German bank—in its role as the trustee of the trust that held the mortgages—issued monthly reports about the performance of the underlying mortgages, essentially a running catalog of the interest and principal payments and whether or not they were being paid on time, if at all. At first, not surprisingly, all went mostly well. In April 2006, 282 of the second mortgages were prepaid voluntarily, leaving a pool of 12,176 mortgages with an outstanding loan amount of $721.9 million. There were no foreclosures by that date, but there was one troubling sign: after just one month, 362 of the mortgages were already delinquent on their payments by at least thirty-one days.

  And then the deterioration began to accelerate. By the end of October 2006, some 799 of the mortgages, or 7.62 percent of the total number, valued at $50.8 million, or 8.3 percent of the total value, were at least thirty-one days delinquent in their payments. By the end of December 2006, 927 mortgages—worth $59.1 million, 10.2 percent of the total—were at least thirty-one days delinquent in their payments. Another 26 homes had been foreclosed upon and another 75 of the borrowers were in bankruptcy proceedings. Some seven months after the mortgages were packaged up by Goldman and sold into the marketplace, 1,029, or 11.2 percent, of the loans were either delinquent, foreclosed upon, or subject to bankruptcy proceedings. The graph of the number of mortgages three months or more behind in their payments looked like a hockey stick and represented 10.3 percent of the total number of loans. By May 2009, according to the lawsuit filed by an investor in GSAMP Trust 2006-S2 against Goldman Sachs (and others), some 12.25 percent of the underlying mortgages in the trust were delinquent. (A subsequent Goldman offering, GSAMP Trust 2006-S3, which followed S2 by only a couple of weeks, had an even worse performance, with some 18 percent of its mortgages in trouble after around nine months, even though the ratings agencies had rated some 73 percent AAA.) Within weeks of the December update, Deutsche Bank had filed a notice “of suspension of duty to file reports” and that was the end of the monthly reports.

  By December 2006, not only had Goldman sold the GSAMP securities—and probably made around $10 million, and likely more, for its trouble (after selling the cakes in the bakery for more than the cost of the ingredients)—but the firm had also made a fundamental decision about the dangers that appeared to be lurking in the mortgage securities market: the time had come for Goldman to get out of the mortgage market.

  CHAPTER 19

  GETTING CLOSER TO HOME

  By the middle of 2006, Josh Birnbaum, and his colleagues Swenny and Deeb Salem, had been buying and selling the ABX index for some six months, mostly just on behalf of the firm’s clients. Birnbaum described this as, “A client says, ‘Where do you bid the index?’ and you give him the price. If he says, ‘Where are you going to sell the index?’ you give a price. Just try to buy as many at lower prices as the ones you sell. In other words, a classic Wall Street market-making model.” (Whatever they were doing seemed to be working. Birnbaum’s desk would make $288 million in the first quarter of 2007, compared with $163 million for the full year in 2006.)

  By around June 2006, Birnbaum started to notice a rise in the number of borrowers becoming delinquent in their mortgages, thanks to the monthly reports filed publicly by the trustees of the mortgage-backed securities such as GSAMP-S2. “There were some early cracks in the subprime market,” he said. “In terms of the early cracks, the way people were looking at stuff back then wasn’t in terms of losses but it was just the borrowers going delinquent.” Prior to that, he said, “many people had qualms and concerns about what was going on in mortgages—and specifically in mortgage credit—but they were more theoretical concerns than corroborated concerns.” Among the concerns being voiced were that borrowers were taking out mortgages they knew they were unlikely to repay, that they didn’t “have any skin in the game,” or that their down payments and FICO scores were too low.

  Hedge-fund manager John Paulson was one of the first to start trading the ABX index aggressively in early 2006, consistent with his utterly bearish macro bet on the U.S. housing market. At first, he did the bulk of his trading through Deutsche Bank. “He was a bit of an enigma to many people at that time,” Birnbaum said, “because he wasn’t anything close to the John Paulson we all know today.” In February or March, Paulson contacted Birnbaum’s desk at Goldman to inquire about trading the index with Goldman as well. “It was one of these things where we get the call on the desk,” Birnbaum recalled, “and it’s Paulson. [People were wondering,] ‘Who is that? Who is this guy? What’s he doing?’ Most people had never met him before. Even the senior guys in the bank. We didn’t know what he was up to.”

  At first, Birnbaum did the ABX index trades for Paulson without giving him too much thought one way or the other. But as the year progressed, Birnbaum started to wonder what the guy was up to and how serious he was planning to be in going short the mortgage market. He asked for a meeting with the Paulson team. He wanted to get to know his client. “It was some combination of a diligence meeting and a kind of ‘testing-the-guy’ meeting,” a Goldman executive said. “There was a concern at Goldman to some extent that ‘did this guy know what he was doing?’ There was almost a kind of a suitability aspect to the meeting because at the time you’re talking about a very different viewpoint about housing and about these indices that are new and that when we look back, it’s a completely different situation. But then we had a suitability issue that we were checking out because he was doing this trade bigger than a lot of folks.” At that time, Paulson was asking Goldman to undertake individual trades of several hundred million dollars at a time. “Which is very significant in the context of—particularly looking back—of what each one of these trades meant in terms of economic value at risk,” he said. In other words, the trades that Paulson started to do with Birnbaum were the very ones that ended up paying off fabulously well for him and his firm, making him a multibillionaire in the process. Goldman needed to make sure Paulson was legit—could satisfy the men in the basement office—before deciding to do serious business with him.

  At the meeting, Birnbaum decided to judge Paulson’s suitability by asking probing questions of Paulson and his team about how much conviction they had about this bet. There was an element of Kabuki theater to the meeting, what with all the alpha-male head fakes. “Part of it was just testing his thesis as to whether or not he knew what he was doing,” Birnbaum said, “and part of this was, frankly, the guy was doing big trades and as a market maker you need to figure out how big is this guy gonna be? And he’s not going to tell you to your face how big he’s gonna be.
He’s not gonna tell you, ‘Oh yeah, I had three more trades to do and they’re gonna be this big,’ because then we might price the market a certain way. He’s gonna try to hold his cards here”—holding his hand close to his chest—“and not reflect what the size of his program is. But as best I could, I was trying to ascertain what’s the size of his program, what’s the strength of his thesis? In doing this, we almost try to play devil’s advocate.”

  Birnbaum was trying to ascertain whether Paulson was going to continue to make trading bets of hundreds of millions of dollars or whether he would be increasing his bets into the billions of dollars, making him potentially a very large and important client. “We were in completely new territory at the time of this product,” he said, “and were just trying to understand his assumptions and poke holes in his assumptions and see how his answers stacked up.” This was a bit of high-stakes psychological warfare. “You can always tell if you poke holes in someone’s assumptions and you see a lot of self-doubt in them. If their assumptions don’t hold up so well, you might have a better sense of what their program is going to look like. On the other hand, if you poke holes in their assumptions and they’re coming back with vehement arguments, you get a different view of what their program is going to look like. A lot of it was just how strongly do these guys really feel when you test them? When you test their thesis that this is going to be the trade of the century … how well does their argument hold up?”

  Trading the ABX index or buying huge amounts of credit-default swaps was not for the faint of heart. It was financial gambling of an unprecedented nature. “It’s not like trading IBM,” Birnbaum said. “These are much riskier transactions. Consider them huge block trading transactions where there’s an element of luck. There’s an element of really needing to understand what are the next three chess moves that your counterparty is intending to do because that’s really going to affect how you price the next trade. It’s just a very different dynamic, and that was even then. Looking back now—where you can see how much the price ended up moving on these things—it’s unprecedented to trade that kind of risk with that frequency on Wall Street.”

  Birnbaum said he thought his intense probing rattled some of the younger professionals on Paulson’s team and got them to question the wisdom of their thinking. “You’ve got a guy who is the main trader … at Goldman Sachs—good firm, reputation and all that—who is sitting there asking probing questions and raising doubt, with the intention of testing their nerve,” he said. “For certain members of their team—and I’m not saying for John [Paulson]—but certain members of their team probably found themselves questioning their assumptions based on the level at which I was asking these questions.” According to The Greatest Trade Ever, Gregory Zuckerman’s bestselling account of how Paulson made his billions, Birnbaum kept calling Brad Rosenberg, Paulson’s trader, to find out how much protection Paulson intended to buy. “When Paulson and [Paolo] Pellegrini”—Paulson’s partner in conceptualizing the bet against the mortgage market—“got wind of Birnbaum’s inquiries, they told Rosenberg to keep him in the dark. They worried that Birnbaum might raise his prices on the CDS insurance if he knew more buying was on the way.”

  According to Zuckerman, Birnbaum persisted with Rosenberg and insisted on having a meeting. “If you want to keep selling, I’ll keep buying,” Birnbaum reportedly told Rosenberg, Paulson, and Pellegrini. “We have a few clients who will take the other side of your trades. And I’ll join them.” Not only was Birnbaum trying to discourage these guys from continuing to make the bearish bets, but he was also “trying to tell Paulson he was making a big mistake.” Rosenberg, in particular, seemed rattled after the meeting with Birnbaum, according to Zuckerman, and walked into Paulson’s office and asked the boss if they should tone things down. “Keep buying, Brad,” Paulson reportedly told Rosenberg. (John Paulson declined repeated requests to be interviewed.) With Paulson’s backing, Rosenberg called Birnbaum when the Goldman trader got back to his desk and told him he wanted to continue the bets against the ABX. “Really?” Birnbaum supposedly said.

  The description of the meeting in Zuckerman’s book was too “heroic,” Birnbaum recalled. “It makes it look like the whole world was against John Paulson—including Goldman Sachs—but that they persevered and they were right,” he said. “What drama! What better bank to personify the world being against Paulson than Goldman, right?” Zuckerman’s account of the meeting was accurate in one sense, he said, in that he, Birnbaum, was trying to intimidate the Paulson team. “Being cynics, they weren’t sure if I was just trying to catch them off their game or maybe was trying to get them to do something different than what they were doing in order to benefit Goldman,” he said.

  Birnbaum said he learned a lot from the meeting. In particular, he said, he came out of it convinced “these guys were here to stay in terms of being active and buying protection on subprime. We dealt with them differently after that.” Goldman ramped up the amount of capital it made available to Paulson & Co. for its huge bets against the ABX index. As important, though, was Birnbaum’s decision after that meeting with Paulson to give deeper thought to Paulson’s trade and its implications. Should Goldman follow suit? Should Birnbaum and his colleagues on the structured products desk get some conviction with regard to what might be happening in the mortgage markets and make a directional bet with Goldman’s capital? Was Paulson … right? Birnbaum’s meeting with Paulson—regardless of which version of it is most accurate—turned out to be momentous, if for no other reason than Goldman soon enough thereafter began mimicking Paulson’s bet. “Not necessarily right at that time, but shortly thereafter in 2006, we were changing our views from being more agnostic to be directionally short,” is the way Birnbaum described this moment.

  With Goldman changing its “view” from being agnostic on the mortgage market to betting aggressively against it, the relationship between Goldman and Paulson became somewhat more adversarial and competitive. Virtually overnight, Paulson & Co. went from being a good client of Goldman Sachs to becoming a competitor of Goldman Sachs. Much to the consternation of many of its clients, it was a dynamic that had occurred often at the firm, especially since Goldman went public in 1999, as it had seriously bulked up its proprietary trading, its hedge-fund business, and its private-equity business and found itself competing with some clients in those businesses. “Obviously if I am looking to short the market, if you’re John Paulson and you’re looking to short the market and you call me up, it’s not my favorite phone call anymore,” Birnbaum explained. “We’ll make a market for you, but we may not be the best price anymore to sell you protection, whereas three or six months earlier you might have been the best price. So there was a change.” But this being Goldman Sachs and Wall Street, Paulson and Goldman would soon enough find other ways to work together, even as they were competing on the purchase of credit-default swaps on the mortgage market or on shorting the ABX index.

  By the beginning of the third quarter of 2006, Birnbaum and his colleagues had decided to get short, real short, and in a big way. Birnbaum is not able to pinpoint the exact moment he made the decision or whether there was one particular reason for his change of heart. It was more a combination of factors. First was the increasing rate of delinquencies in one mortgage-backed security after another. The problems in the GSAMP 2006 deal and its increasing delinquencies were typical of the kind of negative statistics that were catching Birnbaum’s eye. “Whereas the view on the market before that for the Volckers and the Meredith Whitneys and these people was more of a visionary thing: ‘Where this could happen …’ When you actually saw it happening, in the form of delinquencies on some of the more recent subprime vintages—they were showing acceleration of delinquencies that were at a rate where they were at least as bad as the prior bad vintages of 2000, 2001, and in fact they were proving to even be worse as each month went by. The fundamentals were just not improving. The housing market was starting to lose steam.”

  Birnbaum also ha
d at his disposal a proprietary computer model, designed and built by his colleague Jeremy Primer, that allowed him and his team to analyze detailed scenarios of how bad things could get for mortgage securities if the assumptions on defaults were increased be-yond the conventional wisdom. “Pretty complicated stuff to figure out what a CDO is worth,” Birnbaum said. “In common sense, we know it’s not worth one hundred [cents on the dollar], right? But what is it really worth? Is it worth seventy or fifty or whatever?” That’s where Primer and his modeling expertise proved an invaluable competitive advantage. Primer was just your typical Goldman professional. Primer grew up in Maplewood, New Jersey, the son of two English professors at Rutgers University. Naturally, Primer became a math whiz. When he was sixteen, in 1981, he won a gold medal for the USA International Mathematical Olympiad team. A typical problem he had to answer was: Three congruent circles have a common point O and lie inside a given triangle. Each circle touches a pair of sides of the triangle. Prove that the incenter and the circumcenter of the triangle, and the point O are collinear. Primer thought the test was embarrassingly easy. “It was a joke,” he told Time after the test. He graduated magna cum laude from Princeton, was elected to Phi Beta Kappa, and then went on to get a master’s degree in mathematics from Harvard.

  He joined Goldman in 1993 as part of the prepayment modeling group in the mortgage derivatives business. By 2000, he had become head of mortgage modeling and was a prepayment strategist. Primer’s model could analyze all the underlying mortgages and value the cash flows, as well as what would happen if interest rates changed, if prepayments were made, or if the mortgages were refinanced. The model could also spit out a valuation if defaults suddenly spiked upward.

  Primer’s proprietary model was telling Birnbaum that it would not take much to wipe out the value of tranches of a mortgage-backed security that had previously looked very safe, at least in the estimation of the credit-rating agencies that had been paid (by Wall Street) to rate them investment grade. For instance, suppose an investor owned the BBB-rated slice of GSAMP-S2, meaning there were investors supposedly taking more risk that were below the BBB slice in the structure and investors taking less risk in the structure because, as holders of the AAA-rated slice, they had the first claim on incoming cash-flow payments. In a typical mortgage-backed security, 8 percent of it was below BBB rated, 2 percent was BBB rated, and the rest, or 90 percent, was rated higher than BBB. What Birnbaum and Primer figured out was that it would not take a particularly significant increase in mortgage defaults for the investors in the BBB slice of the security to not get paid, even though BBB was an investment-grade rating. “As long as losses are below eight percent, your bond’s money good,” Birnbaum explained, meaning it would be worth 100 cents on the dollar. “You’re fine. If losses are nine percent then you’re losing half on your two percent tranche. If losses are ten percent then you’re losing all of it. So merely by having losses go from eight percent to ten percent you’re going from money good to wiped out. The problem I had with this was that we were basically completely out of sample when it came to predicting where losses would be on these mortgages, particularly if home prices went down. We had no data on it. Okay? So any model that was pretending to know what was going to happen was just completely putting its finger in the air with an arbitrary estimate.”