In a series of recent articles, I have taken readers through the mechanics of each leg of the JPMorgan (JPM) Whale Trade, examined the available data to determine which legs of the trade may still be in play, and discussed with readers what the trade, taken as a whole, says about the corporate culture that exists at JPMorgan. Recently, it was brought to my attention that one detail has been left out of my analysis. After looking into the matter, the issue is both interesting and easy to understand, and serves to reinforce much of what has already been said about the way the firm operates.
First and foremost let me give a brief definition of 'value-at-risk' (VAR). Value-at-risk is a controversial method of estimating the probability of losses by using statistics to analyze the history of relevant prices and volatility. Let me also say that JPMorgan uses a 95% confidence interval (2 sigma) in its VaR models, a practice which all manner of academic research has found to be woefully inadequate.
But leaving that aside, and delving back into a bit of history, consider that in JPMorgan's 2011 third quarter 10Q, the firm has its CIO desk VaR listed at $48 million. Similarly, the firm's 2011 10K says that "CIO VaR averaged $57 million in 2011". All seems well (or at least consistent) at the CIO desk. Then we come to the firm's 10Q for the first quarter of this year, and the CIO VaR jumps to $129 million out of the blue.
Well, at least that's what it looks like to the person looking back on the data now. Unfortunately for the poor soul who read the 10Q when it came out, that same number ($129 million) read just '$67 million', giving the impression that nothing out of the ordinary was taking place over in London. In an unprecedented move, the number was restated later as the firm admitted that the model used to generate the $67 million figure was a new model, has since proven inadequate, and should have been replaced with the old model.
The question however, is not whether the 'new' model is 'inadequate'. The question is whether it was 'convenient'. After all, the new model was put in place on January 13 according to Bloomberg which is precisely the time when Bruno Iksil was beginning to sell massive amounts of protection on the IG9 to hedge his CDX tranche bet which, thanks to the tail-risk crushing LTRO, was going sour fast. The timing here is no coincidence. The CIO desk switched to the new model so they could take the risks they felt they needed to take given the circumstances. In Dimon's own words:
the change in the VaR model and its lower reading "did effectively increase the amount of risk that this unit was able to take,"
It seems abundantly clear that the CIO desk switched to the 'new' model because they knew they would need lower readings to attempt to trade their way out of a mess. The 'new' model gave them half the VaR readings of the old model, and thus allowed the Whale Trade to continue without raising any red flags (for a while).
This is an important part of the story for JPMorgan shareholders, as it shows that the firm most likely switched to a different risk model when the situation began to escalate rather than take the escalation as a sign to be prudent, cut losses, and unwind. This is just another example of how the firm's organizational culture is fraught with arrogance-- a culture no one should be comfortable buying into. As more and more details come to light about the firm, stay long JPMorgan puts, and always beware your risk.