All The Devils Are Here: Unmasking the Men Who Bankrupted the World (46 page)

Which also meant that, like virtually everyone else at Merrill Lynch, Breit had no idea what Semerci was doing with Merrill’s CDO business. Though he was one of the few people left at Merrill with the knowledge and background to sniff out problem trades, he was shut out entirely.

Toward the end of 2006, Merrill Lynch took a final step in its ongoing quest to be the dominant Wall Street player in subprime mortgages. It finally bought that mortgage originator O’Neal had been hankering for. The company was First Franklin, a unit of National City Corporation that had made $29 billion in mortgage loans the year before, virtually all of them subprime.
^*
The purchase price was $1.3 billion. Now Merrill would have its own source of mortgages that it could securitize to its heart’s content. Or so the company hoped.

The Merrill executive who had been handed the job of landing a mortgage company for Merrill was Michael Blum. In truth, Blum was not a big fan of the First Franklin purchase, though he had dutifully completed it. In 2005, he and his staff held a meeting with O’Neal to lay out Merrill’s possible options for getting into the mortgage origination business. O’Neal was eager to get going; Lehman Brothers already made four times what Merrill made in mortgages, in part because it owned BNC Mortgage, the eighth largest subprime company in the country. Blum thought the firm should start up its own originator rather than buy one. “Buying something will be painful because they are not well-managed companies and they are at the bottom of the food chain,” he told O’Neal, according to people who were in the meeting.

O’Neal asked him how long it would take to build a subprime company
from the ground up. Three or four years, replied Blum. O’Neal gave Blum a steely look. “I’m fifty-four fucking years old,” he replied, “and I don’t have three or four years.”

Blum was also astounded by the price Merrill was willing to pay. (After the deal was announced, David Daberko, National City’s CEO, told Dow Jones News Service that the deal “is exactly what we were hoping for.”) But Merrill was so eager to get in the game that it would likely have paid even more; Blum could take comfort only in the fact that it wasn’t New Century. First Franklin was supposed to be one of the better-run subprime companies.

Almost immediately after the deal was completed, First Franklin began taking losses. Like all the subprime originators, it had kept the residuals and posted gains that reflected an optimistic estimate of their value. Now, as delinquencies rose, those gains were being reversed. In a meeting in January 2007, as Blum was going through the losses in the residuals book, Dow Kim suddenly looked up from his BlackBerry with some news. “Lehman Brothers just had a record quarter in mortgages,” he said, according to someone at the meeting. “I guess they’re just smarter than we are,” Blum replied.

And so it went. In a February meeting, Blum and his team argued that delinquencies were likely to get worse. He wanted Merrill to start hedging its exposure. Dale Lattanzio, whom Semerci had installed to run the CDO business after Ricciardi, brought out a series of charts, using the 1998 Long-Term Capital Management failure as his worst-case scenario. If something like that were to happen, he said, Merrill could lose as much as $70 million. Anything short of that scenario, the firm would be fine. (Lattanzio denies the details of this meeting, which were recounted by several people who were there, and says that he did not list $70 million as a worst-cast scenario.)

After the meeting, a risk manager told Kim that “there’s no way” Lattanzio’s estimate was right. Kim asked the risk manager to poke around and come up with a better estimate, according to a former Merrill executive. But the risk manager couldn’t get any information out of Lattanzio and Semerci, and had to drop the effort.

Blum couldn’t understand how the people running the CDO business could be so sanguine. They were using the same raw material he was: subprime mortgages. By early 2007, defaults were the highest they had been in six years, when subprime one had collapsed. HSBC, the big British bank, said in February that its bad debt charges would be 20 percent higher than previously anticipated, thanks to its deteriorating subprime business. And yet Merrill’s mortgage desk continued to churn out CDOs and post profits.
In the first quarter of 2007, the firm underwrote twenty-six CDOs, of which nineteen were made up primarily of subprime mortgages. First Franklin was taking $50 million to $100 million in quarterly write-downs, and top management at Merrill was all over Blum about its deteriorating financials. Yet somehow the CDO business remained untouched. How could this be?

In April, Blum gave a presentation to the board in which he put forth a downbeat and sober-minded assessment of the subprime business. Afterward, many of the board members sent him thank-you e-mails for his plainspoken presentation. What they had failed to notice, however, was that seated next to Blum at the board meeting was Osman Semerci. He never said a word about any problems he might be having with subprime mortgages. Nor did anyone think to ask him.

All over Wall Street, an immense amount of risk was building up in the system. It wasn’t just that firms were taking on risk when they bought subprime mortgages and bundled them into securities, or when they kept some of the leftover pieces themselves, or when they bought whole subprime mortgage originators. Over the course of a decade, subprime mortgages had managed to seep into Wall Street’s bloodstream, as firms used products created out of them to increase leverage, reduce capital, generate profits, and, more generally, game the risk-based rules that were originally intended to give firms the flexibility to deal with the modern world. All of which also meant that the increasing risk was masked by layer upon layer of complexity, hidden where few on the outside could see it.

For instance, using a loophole in Basel I, banks set up off-balance-sheet entities that came to be known as SIVs, or structured investment vehicles. In a nutshell, banks didn’t have to hold any capital—that’s right,
no capital
—against these vehicles as long as their outstanding debt had a term of less than a year. (That’s part of why Karen Shaw Petrou, the managing partner at Federal Financial Analytics, says: “Nothing about this crisis was in fact unforeseen. It was just unaddressed.”) By the summer of 2007, there were twenty-nine SIVs with outstanding debt totaling $368 billion, of which nearly $100 billion belonged to Citigroup-sponsored SIVs. Because SIVs were looking for yield, just like every other buyer of triple-A securities, many of them began to buy more and more mortgage-backed securities. Ostensibly, SIVs were independent from the sponsoring bank. But if there was a crisis and the
debt started to default, would an institution like Citigroup really be able to sit back and let the SIVs fail? Or would it have to rush in and put that debt on its own balance sheet, which would have a crippling effect on its capital?

Another source of hidden risk was in the plumbing of the market—plumbing that was utterly taken for granted. The big banks all had warehouse lines that the mortgage originators borrowed against to make their subprime loans. It was the primary funding mechanism for the industry. But the banks didn’t just extend a big loan to the originators. Instead, they had discovered a more modern, efficient, capital-gaming way to do it. They would set up an off-balance-sheet vehicle that issued short-term commercial paper to fund itself. That commercial paper was backed by the mortgages. It was part of a market called ABCP, or asset-backed commercial paper. According to Fitch, by the spring of 2007 this market was shockingly big: $1.4 trillion in size. The commercial paper got a top rating from the rating agencies, making it possible for money market funds to buy it. However, in order to obtain that all-important top rating, the sponsoring bank, or another bank, invariably had to provide some kind of guarantee, in the event that the vehicle found itself unable to replace the commercial paper when it came due.

As the market got crazier, money market funds became more and more enamored of this paper; they, too, were competing for that extra little bit of yield. Although money market funds were serving the role of the old-fashioned bank—they were ultimately the real lender—they weren’t regulated the way banks were. Since they were holding highly rated securities—as SEC rules required them to do—no one in the government was concerned with the quality of the collateral.

But what would happen if the money market funds all started questioning the quality of the assets backing their paper at the same time? What if they all stopped buying it? Either the sponsoring bank would have to provide liquidity—damaging its own balance sheet—or the vehicles would all have to start dumping assets to raise cash. Neither scenario was pleasant to contemplate.

Money market funds were also a core enabler of the deepest, darkest, least noticed part of the market’s plumbing. This was the so-called repo market, which made it possible for firms to pledge assets in return for extremely short-term loans, often as short as overnight. Yale economist Gary Gorton—the game man who did risk modeling for AIG-FP—explains the repo market this way: Suppose Fidelity has $500 million in cash that it plans to use to eventually buy securities. It wants a safe place to earn interest on that cash while making sure the money will be available the instant it wants it back.
Enter the repo market. Fidelity can deposit the $500 million with an investment bank—Bear Stearns, in Gorton’s example—and be sure the money is safe, because Bear provides collateral to back up the loan. The difference between the money Fidelity gives Bear and the value of its collateral is called the “haircut,” and before the crisis a 2 percent haircut—meaning Bear could get 98 cents in cash for every $1 in assets it pledged—was a normal number.

Secured lending, or lending against collateral, is almost always less risky than unsecured lending. On the Street, the repo market is called the last line of defense, because you can get money there when you can’t get it anywhere else.

And yet, there were dangers in the repo market, too. It is a murky market, but a huge one: according to a report by the Bank for International Settlements, by 2007 the U.S. investment banks funded roughly half of their assets using the repo market. For firms that depended on this market, there could be a timing mismatch, because banks could pledge a long-term illiquid asset in return for short-term funding. If the short-term funds went away, they still had the asset—which needed financing. Another danger was that repo transactions are exempt from the normal bankruptcy process. Lenders didn’t have to worry about their money getting tied up—they could simply grab their collateral at the first sign of weakness. And whichever lender grabbed first did best: no bankruptcy court judge was going to come along and decide what was and wasn’t fair.

As the bubble grew, Street firms began using riskier and riskier assets—including mortgage-backed securities—as repo collateral. They did it for the usual reason: the lender could get a bigger return by accepting mortgage-backed securities as collateral than it could by accepting Treasuries.

But once again, what would happen if the lenders began to question the true value of the collateral? The lender might demand a bigger haircut—meaning that the loan the bank would get would shrink, and it would have to rapidly sell assets, or face a shortage of funds. Or what if the lender didn’t want any collateral from a particular firm at all? Suddenly a routine repo transaction would be transformed into something far more ominous: a vote on whether an investment bank should survive.

Thanks to deposit insurance, the days were long gone when bank customers stood in line to pull their money out of a shaky bank, creating a run on the bank that usually ended in its collapse. But as Gorton and fellow Yale economist Andrew Metrick would later argue in a paper, the repo market created the conditions for the modern version of the bank run. You never saw this kind of bank run in photographs, but it was every bit as devastating.