JPMorgan CEO Jamie Dimon may not want to discuss it during his testimony to the US Senate Banking Committee Wednesday but the $2bn loss wracked-up by the so-called "London Whale" has re-ignited the industry debate over one of its key risk management theories.
Designed to give banks a single, simple figure of their total risk exposure, Value-at-Risk models of risk management, known in the industry as VaR, are being challenged by many experts who say too many banks rely on what's become an out-dated, misunderstood and poorly-used method of calculation in a financial market universe that has fundamentally changed since the collapse of Lehman Brothers in 2008.
"A fundamental problem with the use of VaR, and other quantitative models, in my opinion, is misunderstanding and misuse," says Shannon Burchett, managing director of Risk Limited and one of the most respected figures on the subject. "One of the only positive things I suppose can be said about shocking losses occurring when people have been somewhat comfortably relying on VaR as a sure indicator of expected maximum risk exposures, is that market confidence has now been shaken and risks measurements and risk controls are going to be reassessed."
The losses, indeed, have been severe. Bruno Iksil, nicknamed the London Whale for his rumoured preference for large trades, and his colleagues at the London unit of JPMorgan's "Chief Investment Office (CIO)" are said to have lost around $2bn last month through bad bets in a portfolio that was specifically designed to hedge the bank's risk exposure.
The market reaction to JPMorgan's share price since the loss was revealed has cost investors more than $26.5bn in market value.
This prompted an investigation by several US authorities and Dimon will now have to explain why he didn't ensure that the CIO's risk managers adequately keep pace with the nature of the unit's business.
In prepared testimony for his appearance before US lawmakers, Dimon says the portfolio "... morphed into something that, rather than protect the firm, created new and potentially larger risks. We have let a lot of people down, and we are sorry for it."
What he didn't mention, however, was VaR. Or, perhaps more specifically, the changes made to JPMorgan's VaR calculations, the timing of which are part of an on-going Securities and Exchange Commission investigation.
Misuse and Flawed Use of VaR
One of the key criticisms Dimon, JP Morgan and its CIO face is the flawed, inadequate or misuse of VaR, which is a broad calculation model, used by trading firms and investment banks to measure the potential losses of a portfolio.
Virtually all investment banks use this as part of their risk management practices in order to prevent substantial swings in profits and losses. However, implementing such a measure is only an indicator of risk and does not necessarily dictate trading actions.
Dimon has infamously dismissed the use of VaR in 2006, he said it was "a very bad number if you think it actually represents risk". A few years later, in 2009, he said he didn't "pay that much attention to VaR."
Given the whale-like losses at the London CIO, it may seem like a clear cut case in terms of "dismissing VaR at your peril." However, the case is not black or white.
JPMorgan changed its some of the calculations and measurements in its VaR model this year, and indicated that the average Value-at-risk figure by $2m to $67m for JP Morgan's CIO unit.
However, reports suggested in April that the CIO's portfolio was so large that is was distorting market prices. Dimon shrugged-off the suggestion as a "tempest in a teapot". The following month, he said the CIO loss of at least $2bn and announced a return to the previous VaR formula, which revealed that average figure had in fact almost doubled to $129bn.
Time to Move Faraway from VaR?
The JP Morgan case is not new, in terms of the flawed or misuse of VaR, but it does highlight the critical situation banks and firms can be faced with when risk management goes wrong.
"VaR popularity and the concept have definitely been tarnished by recent high profile losses," says Risk Limited's Burchett, who once worked for JPMorgan, although not directy in its risk metrics group. "Losses by the firm that is considered to have originated the concepts and primary methodologies that lead to VaR, has particularly caught the market attention. Probability analysis is not going away, but it is going to be intensely scrutinized and will also be supplemented."
That's exactly the view of Aite Group analyst Virginie O'Shea, who specialises in what's known as post trade technology, who argues that VaR should be only one of many tools banks employ to understand, measure and better control risk.
"I don't think VaR is outdated but I think the problem is when a group has an over reliance on VaR and using it as the only model that they based risk calculations on," she says. "VaR is designed to be part of a wider toolkit. VaR still has a place but just as part of a bigger picture. The models are never accurate but they should be used as more of an indicator."
The US Office of the Comptroller of the Currency (OCC) similarly warned banks to scrutinise possible VaR flaws in April last year and this year said that the losses reflect JP Morgan's CIO's "inadequate risk management".
So do Dimon's comments, the OCC's warning and the huge losses suggest a confused relationship between the users and the consumers of VaR?
Knowledge is power
"I suppose the short answer to whether VaR have a place in financial risk management anymore, yes somewhat, but only as one of various risk management tools and much of this is diminished as the universal measure and any view of it being an absolute or infallible risk measure is gone," says Risk Limited's Burchett.
The banks themselves have been addressing this issue, including JPMorgan, whose analysts have published papers on VaR's use to assess the market risk of a portfolio of traditional financial instruments.
In the sixth article in a series, Romain Berry who works in the JP Morgan Investment Analytics & Consulting unit at the bank, said "we explained at a high level the pros and cons of the three main methodologies, namely Analytical, Historical and Monte Carlo Simulations [VaR models]."
"In our last article we described how one can stress-test a portfolio to perform some sensitivity analysis or to account for extreme movements in the markets. But how accurate are these various VaR methodologies? At the end of the day, all these models may look elegant and quite sophisticated, but do they really do what they say they do - identify some areas in a portfolio where risks may arise and eventually affect the overall performance of the portfolio?"
"Reliance on statistical methods has it limits," says Burchett. "VaR had/has the appeal of an easy, all inclusive answer, which sort of was the original point of the model and has been in vogue and universal for some time, even without a lot of understanding by executive users in some instances. However, we are in a period of high instability, in virtually every market segment, which tests forecasting and valuation models; and the overall financial crisis and specific firm mishaps and losses have obviously highlighted the quantitative model failures on a large scale."
Adding components to the risk management model, however, won't happen overnight, says Aite Group's O'Shea.
"I think it will take time as many institutions have VaR imbedded in their processes but people are using other methods to calculate expected loss as Basel III will force stress testing and new Financial Services Authority rules will forces groups to use tools to assess liquidity risk, which VaR cannot physically do. The more data you have, the more powerful you are."
By the time of publication, JP Morgan had not responded to calls for comment on this article.