Risk Management

The Hedge That Wasn’t

JPMorgan Chase’s $2 billion trading miscue is a costly lesson in how not to protect against potential losses.
Vincent RyanAugust 1, 2012

The $2 billion trading loss at JPMorgan Chase that came to light in May stemmed from fundamental mistakes by risk managers — in particular, the belief that a hedge on credit exposures could both reduce the bank’s risk and earn billions of dollars at the same time.

That’s what experts are saying about a trading loss that knocked more than $20 billion off the bank’s market value, sparked probes by the Justice Department and the Securities and Exchange Commission, forced credit-rating agencies to issue negative outlooks for the bank, and turned the spotlight on a bank unit that was set up to invest excess deposits but also generate a sizable profit.

According to JPMorgan, the bank’s chief investment office was investing in a benchmark for credit-default swaps designed to mitigate the bank’s overall credit exposure, especially the possibility of higher interest rates and inflation. But the hedge was risky in itself, and the positions in derivatives were so large that they distorted the market, experts say.

JPMorgan also may have failed to fully understand its exposure because it was relying too heavily on value at risk (VaR), a common risk model that estimates the potential loss in value of a risky asset or portfolio over a certain number of trading days.

In May, JPMorgan stated that the trading positions — reportedly on a series of the 10-year Markit CDX North American Investment Grade index — were “riskier, more volatile, and less effective as an economic hedge than the firm previously believed.” JPMorgan CEO Jamie Dimon admitted that they were “flawed, complex, poorly reviewed, poorly executed, and poorly monitored.” Later that month at the bank’s annual shareholder meeting in Tampa, Dimon added that he couldn’t justify the mistake.

Improper from the Start
“It screamed improper credit-risk management to us from the very beginning,” says Matthew Streeter, a product manager at FINCAD, speaking of when he heard about the incident. (FINCAD is a developer of derivatives risk-management software.)

For one thing, the credit-index bets introduced secondary risk exposures that JPMorgan executives apparently didn’t take into account, says Streeter. A key question to ask about hedging, he says, is whether a hedge exposes the insured to risk above and beyond the risk the hedge is supposed to mitigate.

For another, the trades JPMorgan was making were so large that they were moving the market for the hedging instrument, concentrating a lot of risk on one side of the trade. “Any hedge on a $650 billion book will absolutely drive the market,” Streeter says.

Also, JPMorgan’s chief investment office was doing what Streeter calls “speculative hedging”: hedging with the intention of making a profit. As Ryan Gibbons, managing partner of GPS Capital Markets, told CFO two years ago, big gains and losses from hedging spell trouble. When companies make money from hedging, they tend to forget what their core businesses are, said Gibbons. At the high point of the chief investment office’s performance, in 2009, it managed nearly 17% of the bank’s assets and generated $3.1 billion in yearly income.

But a former finance chief at JPMorgan takes the stand that hedging is inherently risky, and that if the firm was taking a risk, it should have expected to get paid for it. “Who says you don’t earn money on a hedge?” Dina Dublon, finance chief of JPMorgan Chase from 1998 to 2004, asked CFO in May. “The only time hedging is not risk-taking is when you sell the position you are trying to hedge; anything else, you’re taking risk.”

Dublon said JPMorgan’s chief investment office was not a separate entity when she was finance chief, but was housed within the bank’s treasury function. Whether or not her finance department hedged credit risk depended on the price of the hedge, she said. “Sometimes you do, sometimes you don’t.”

What Value at Risk?
When JPMorgan revealed the loss on May 10, it said its risk measure had underestimated the portfolio’s volatility. In the first quarter, JPMorgan was monitoring its positions using a new VaR model, which was indicating the chief investment office unit value at risk was $67 million. But JPMorgan scrapped that model and returned to an older version once the loss was discovered. It restated the VaR for its first quarter as $129 million.

“A VaR of $67 million at a 95% confidence level implies the [chief investment office] should not incur losses of more than $67 million on more than three business days during the quarter,” wrote CreditSights analyst David Hendler in a report. But the restatement meant that the office “may have incurred losses in excess of $129 million on three days.” Hendler concluded that the losses put in question not only the accuracy of JPMorgan’s VaR models but also the data being input into those models.

A March 2008 report by a group of senior global-banking regulators, “Observations on Risk Management Practices During the Recent Market Turbulence,” noted that many banks used VaR, but did so in conjunction with notional limits, stress tests, and forward-looking scenario analysis. “Firms that avoided significant unexpected losses used a wide range of risk measures to discuss and challenge views on credit and market risk,” said the report.

Despite all the tools available, regulators and large, complex banking organizations face a tough task in judging the riskiness of many nontraditional commercial-banking activities, testified Federal Deposit Insurance Corp. vice chair Thomas Hoenig to a Senate committee in May. “Trading and market-making are high-frequency activities that can take place between [regulator] exams with little evidence that they ever occurred,” he stated; monitoring trading on a “high-frequency basis” would be costly for banks and regulators.

Hoenig is proposing a strict division of some of the banking functions conducted by the big universal banks like JPMorgan. Under his proposal, commercial banks would be allowed to perform some investment-banking activities but would be prevented from making markets in derivatives or securities and trading securities or derivatives for their own account or a customer’s.

But there is little enthusiasm on Wall Street for such an arrangement. Asked by a Senate committee in June to comment about whether the Dodd-Frank Act’s so-called Volcker Rule, which would limit banks’ proprietary trading, could have prevented JPMorgan’s loss, CEO Dimon replied that the rule was “unnecessary.”