Wall Street Turns on Its Clients

The first quarter of Goldman's 2007 financial year, which ran from the end of November to the end of February, was pivotal for the mortgage team. In December and January, as the mezzanine floors of the ABX became more and more inundated, the traders were changing their views. It was getting harder and harder to sell long positions in the market. Birnbaum, Lehman, and Swenson began to suspect that Paulson was right and that they needed to get into the “big short” while there was still time.

Meanwhile, down at the level of subprime originators, the virus of early payment defaults that Gasvoda had first detected in late 2006 was spreading to the warehouse of mortgage loans that the originators had transferred to Goldman after advancing cash to home owners. Because of the problems Gasvoda was having, Sparks had started to write down the market value of the originators' mortgages, and that meant asking for additional margin to cover the difference. But when Sparks made his margin call, the originators couldn't pay. As Gasvoda recounted to Sparks, “We just sent our guys in to collect some files. And they kicked our employees out of their building with a security guard.” Sparks and Gasvoda began to realize that the same thing was happening at all the Wall Street firms — they were demanding their money in buybacks or margin calls, and as a result the mortgage originators were dying like flies. Sparks was left with subprime mortgages he couldn't sell and loans that weren't being repaid.

As he attended the weekly risk meetings with the heads of nine other key Goldman risk-taking businesses, Sparks felt himself once again under attack as he explained how the mortgage pipeline had unexpectedly malfunctioned. Particularly attentive was Craig Broderick, the firm's head of credit risk management, who was equally perplexed that the client margin calls had suddenly turned ugly. “We're not taking big risk positions here,” said Broderick. “We're just acting as an intermediary on a really safe basis earning relatively low returns.”

Goldman considered itself to have world-class risk management, but here was something their risk managers had completely missed. How could a mortgage pipeline business that Broderick characterized as “pretty low-volatility, pretty highly liquid, pretty customer franchise-oriented” misbehave in this way? “We haven't been modelling this stuff well,” Sparks told him. “Our risk metrics are really, really bad.” Throughout 2007, Sparks's mortgage traders would repeatedly clash with Goldman risk managers over the signals being given by the firm's risk management systems.

Fed by Birnbaum's growing short positions, Pete Ostrem's CDO business may have boomed during the first quarter of 2007, just like its competitors, but Sparks was focusing on the warehouse of unsold mortgage loans and bonds. He and the firm's financial controllers were now updating the value of Ostrem's and Gasvoda's warehouses using the information they were getting from the mortgage bond market, as well as derivatives including the ABX index. That was an innovation that terrified subprime originators like Fremont or New Century — when used to calculate margin calls by Wall Street firms, the new derivatives radar system was literally putting them out of business. Seeing everything holistically — “these are all just positions now,” Sparks would tell his team — meant that Goldman saw the world in a new way, differently from its clients. It was also radically different from the way the rest of Wall Street did things. The giants of the subprime CDO market — Merrill Lynch, Citigroup, and UBS — had the biggest warehouses of all, but persisted in ignoring what the radar system was saying.

Does it matter if a Wall Street firm changes its industrial process from something recognizable — buying raw materials, warehousing them, and repackaging for distribution — to something based on derivatives, where it is taking a proprietary position against its clients? The end products look almost the same, but the motives of the manufacturer can be very different. With the warehouse model, the traditionalists like Merrill or Citi had to eat their own cooking, to the extent that demand might suddenly fall away, perhaps due to a product quality issue, leaving the wholesaler stuck with unsold inventory. However, with a derivative contract, the buyer and seller remain connected, and if the buyer loses, the seller has to win. When Goldman began taking short positions against the buyers of its Abacus CDOs in 2004, the conflict wasn't an issue. The size of Goldman's traditional mortgage warehouse, along with its long positions in derivatives, far exceeded the Abacus positions, so on a net basis, Goldman was indeed eating its own cooking.

In late 2006, this changed. Alerted by the signal of the ABX, Sparks's traders realized that short positions were not just bookkeeping, but a proprietary bet they needed to win. Over a nine-month period from November, Goldman allowed its warehouse of almost $8 billion in unpackaged subprime loans and $7 billion in mortgage bonds destined for CDOs to gradually run down to $3 billion, while Sparks's traders accumulated new short derivative positions that went as high as $12 billion. In fact, from February until December 2007, Sparks's traders were never less than $2 billion short subprime.

 Although it was not until May 2007 that Sparks officially pulled the plug on Ostrem's warehouse-driven CDO factory — telling his team “I don't care, shut 'em all down” and handing the positions to Birnbaum — from December 2006 onwards, anyone buying a subprime CDO from Goldman was unwittingly betting against the firm's superior insight and market knowledge. While investors like IKB are described as “sophisticated,” the truth is that they were anchored in the cash warehouse world, where Wall Street was as exposed as they were. In testimony to Congress, and in private interviews, Goldman officials insist that the firm did not conspire to bet against its clients. But by imperceptibly changing from a cash warehouse to short derivatives positions — which it knew was more than mere bookkeeping — Goldman made itself look evil and Machiavellian.

None of this was remotely illegal in what was an unregulated market — the small print of derivative contracts points out that a dealer owes no fiduciary duty to the investor and may be actively betting against them. And in defense of Goldman, it was now doing what Deutsche Bank had been doing since 2005 — matching up hate-to-lose CDO investors on one side with Greg Lippmann's proprietary bets or bearish hedge fund managers on the other, such as John Paulson and Elliott Associates.

Lippmann's multiple roles as market-making middleman, virtual CDO manufacturer, and Paulson-style high-roller who stood to benefit from the meltdown of products that Deutsche and other banks created would be enough to make any observer of Wall Street queasy. And the way Deutsche's sales force peddled Lippmann's virtual CDOs to less-sophisticated institutions across the world shows financial innovation at its socially harmful worst. Yet it would be unfair to pin the blame on Lippmann. Well before the crisis broke, Lippmann was open about shorting subprime to the extent of annoying his colleagues who still believed in the U.S. housing market. They resented the premium he paid for protection, which cut into their bonus pool. Lippmann even offered subprime derivatives protection to LB Kiel (by then re-named HSH Nordbank) in 2006, but the Germans were unable to get board approval for the transaction, which would have saved them $3 billion. In contrast with Lippmann and his openness, Goldman's secret last minute 180-degree turn is all the more shocking.

Ironically, it was Goldman's belated attempts to follow in Deutsche's footsteps and match up subprime bulls and bears that eventually landed the firm with a $550 million SEC fine. At Goldman, a key figure was Fabrice Tourre, a young French derivatives marketing wizard who had been posted to New York from Europe to work with Jonathan Egol. With Viniar's orders to “get closer to home” ringing in his ears, Sparks had to find a mechanism to stop his growing short positions from getting too big. “Fab,” as everyone called him, had a marketing pitch: he proposed “renting” the Abacus platform to hedge funds like Paulson & Co.

After overcoming opposition from Sparks's traders, who wanted to earmark the Abacus platform for their own short-selling, in March 2007 Tourre arranged a synthetic CDO called Abacus 2007 AC-1 for the purpose of allowing Paulson to short a billion dollars worth of subprime. The detail that landed Goldman and Tourre in the SEC's gunsights was a minor tweak to the static Abacus template: Tourre brought in a monoline insurance company called ACA to select the specific subprime bonds in the CDO while allowing Paulson extensive influence over the selection. ACA, which via its monoline entity took on $850 million of super-senior exposure, seemed unaware of Paulson's true role, and so was IKB, which bought $150 million of triple-A-rated Abacus notes. In his personal e-mails, Tourre couldn't help referring to the uselessness of all this innovation. “This product is a creation of pure intellectual masturbation,” he confided to a friend, unwittingly writing the epitaph of an entire industry.

 The irony is that had Goldman matched up Paulson and IKB directly in the style of Deutsche Bank, or allowed Birnbaum to take the other side of IKB's subprime bet via an Abacus CDO, the firm would probably have stayed out of trouble.

The full-scale meltdown of the financial system was now only a couple of months away. To understand why it took so long to happen, why Goldman and its rivals were able to keep selling subprime CDOs as long as they did, we need to ask the following question: why did investors like IKB not pay closer attention to what they were doing? The answer is that they thought they didn't need to.

Add comment


Security code
Refresh

2018 © Copyright SuccessMill.com | Finance and Investment