Breaking News

Banco Espirito Santo Drops as Much as 41.5% in Lisbon Trading
Tweet TWEET

JPMorgan Deploys New Risk Model for Derivative Bet

JPMorgan Chase & Co. (JPM), whose trading loss of more than $6.2 billion was fueled by the adoption of a flawed mathematical formula that understated the risks, is trying yet another one.

JPMorgan said today it started using a new formula to judge the risk of the derivatives position, at least the third such model it’s used this year, when it moved most of the contracts to the investment-bank unit. The new analysis cut the firm’s calculation of overall value-at-risk, or VaR, by $36 million, or 24 percent, to $115 million in the third quarter, the New York- based bank said today on its website.

JPMorgan’s switch to a new risk model in January may have helped fuel the trading loss at the chief investment office, Chief Executive Officer Jamie Dimon told the Senate Banking Committee in June. Dimon said May 10 that the bank had reviewed the effectiveness of that VaR model, deemed it “inadequate” and decided to return to the previous version. Restoring the use of the earlier model meant the risk was twice what the bank had reported in April.

“VaR models change almost every time we talk,” Dimon said today in a call with journalists. “When we moved it to the investment bank, they adopted, particularly for the synthetic credit portfolio -- and there are some other changes too -- the investment bank’s model, which we think was the best one.”

The January switch and the timing of the firm’s disclosures are the focus of an inquiry by the Securities and Exchange Commission as the U.S. examines how long executives knew about the CIO’s swelling bets and losses. VaR is an internal estimate of the maximum a portfolio could lose on 95 percent of days, or the minimum it could lose on 5 percent of days.

‘More Frequent’

“These kinds of changes seem to be a little more frequent than what I would normally be used to seeing,” said Cliff Rossi, a former chief risk officer for Citigroup Inc.’s Consumer Lending Group. The Office of the Comptroller of the Currency “will be wanting to take a much closer look at the validity and the stability of that model and its underlying assumptions to make sure that what they’ve got in there now is better than what they had in the last two iterations.”

Bryan Hubbard, an OCC spokesman, didn’t immediately respond to a request for comment.

The firm said today that the investment bank unit’s position assumed from the CIO had a “modest” loss in the third quarter. The CIO had a loss of about $449 million in the period as it “effectively closed out” its position.

‘Still Unaware’

The model that was implemented in January “did effectively increase the amount of risk this unit was able to take,” Dimon told the Senate panel in June. On April 13, when he downplayed the risks of trades on a call with analysts, “we were still unaware that the model might have contributed to the problem,” Dimon testified. “So when we found out later on, we went back to the old model.”

Dimon told lawmakers in June that the models “never are totally accurate in capturing changes in business, concentration, liquidity or geopolitics.” JPMorgan uses data going back one year, Dimon has said, so the VaR figure might fall just because a volatile week was no longer included in the calculation.

“This is all nonsense,” said Steve Allen, former head of risk methodology for JPMorgan who retired in 2004. “Because we know that the risk positions are illiquid, we know with certainty that these numbers are irrelevant to the risk of the position.”

‘Another Example’

JPMorgan has added a warning in its most recent quarterly report that risk models are continually tweaked to account for “improvements” in modeling techniques, and SEC Chairman Mary Schapiro has publicly asserted that banks should disclose significant changes and the reasons.

“It’s another example of the fact that banks are very opaque, they’ve always been opaque, they continue to be opaque, and running around spending a lot of time analyzing changes in the little bit of data that they give out is not really contributing very much to the knowledge base of the world,” Allen said.

While models must be kept up to date to reflect market conditions, too many significant changes may destroy their usefulness, said Rossi, who is now executive-in-residence at the Robert H. Smith School of Business at the University of Maryland.

“These are important benchmarks by which you’re setting risk tolerance in your trading book,” said Rossi. “If things are shifting around, it’s like having a compass that’s spinning around because you can’t anchor it to something. That’s problematic from my standpoint.”

To contact the reporters on this story: Michael J. Moore in New York at mmoore55@bloomberg.net; Dawn Kopecki in New York at dkopecki@bloomberg.net; Bradley Keoun in New York at bkeoun@bloomberg.net

To contact the editor responsible for this story: David Scheer at dscheer@bloomberg.net

Bloomberg reserves the right to remove comments but is under no obligation to do so, or to explain individual moderation decisions.

Please enable JavaScript to view the comments powered by Disqus.