All the Devils Are Here [147]
There was one person at Merrill Lynch who might well have asked the right questions, had he been in a position to do so. His name was John Breit, and he was a risk manager who specialized in evaluating derivatives risk. A calm, soft-spoken ex-physicist, Breit had joined the long march from academia to Wall Street, landing at Merrill Lynch in 1990. He was hardly antiderivatives; like most quants, he believed that derivatives were a useful tool. Nor was he the kind of risk manager who feared all risk. On the contrary, he was one of the people who believed that Merrill had shot itself in the foot by being too risk averse in years gone by.
The problem with O’Neal’s Merrill, Breit believed, was that even as the CEO was pushing the desks to take more risk, the institution still recoiled at a $50 million loss. Merrill’s schizophrenia about risk caused traders to seek out risks that wouldn’t show up in the risk models, Breit believed. “If the VaR is small,” he liked to say, “it means we are taking risk in things we can’t measure.” Breit used to tell Merrill management that VaR didn’t measure black swans—the rare but real risks that could destroy a firm. That was its fatal flaw.
Breit had also learned over the years that by the standards of a physicist, Wall Street was quantitatively illiterate. Executives learned terms like “standard deviation” and “normal distribution,” but they didn’t really understand the math, so they got lulled into thinking it was magic. Traders came to believe the formulas were not an approximation of reality but reality itself. Which is also why firms needed good risk management departments, he believed. The risk managers were the ones who imposed the reality checks that the traders preferred to ignore.
But ever since Fakahany had been put in charge of it, Merrill Lynch’s risk department had been in steep decline. Historically, the top risk executives at Merrill reported directly to the chief financial officer. That was fine when Merrill had a strong CFO like Tom Patrick, who knew that part of the job was to adjudicate the inevitable disputes between the risk managers and the trading desks over what constituted too much risk. As head of Merrill’s market risk, Breit had reported to Patrick.
Once Fakahany took over risk management, the risk officers’ influence began to wane. Within a year, Breit lost his access to the board. Fakahany seemed to view the disputes between traders and risk managers as “squabbling among children,” as a former risk manager put it. Slowly, risk management went from being primarily a front office function—meaning that risk managers sat on trading desks—to a back office function, where they looked at models and spreadsheets and had very little interaction with the traders. “And they started to make less money,” a former risk manager explains. A number of good risk managers either left Merrill Lynch or became traders.
In early 2005, Fakahany decided to push the risk management function down a notch further. He promoted the executive who was head of credit risk to be a kind of risk czar, to whom all the other risk managers would now have to report, instead of to Fakahany directly. Furious at what he saw as the degradation of the risk function, Breit sent Merrill’s CFO, Jeff Edwards, a letter of resignation and he left the firm.
He was away for only a few months, however. Late that spring, one of the fixed-income desks suffered a big loss. Kim tracked down Breit and asked him to return to Merrill, where he would have a desk on the trading floor and work for him personally. Although Breit rejoined the firm’s risk oversight committee, he had no real authority within the firm. Because Kronthal and the other veteran traders knew him and trusted him, Breit was able to develop what he called his “spy network,” to keep apprised of the risks the desks were taking. Once Semerci took over, everything changed. The spy network dried up. Dow Kim, who wanted to leave and start a hedge fund—O’Neal had asked him to stay after the Kronthal firing—was losing interest. (He would