London Whale Swims Off Into The Sunset

Author:
Updated:
Original:

Hi Whale! I told you you were not forgotten. Not understood, either, but not forgotten.

The London Whale now goes by the less adorable name "synthetic credit portfolio," since all mammalian representatives of that portfolio have left for non-extradition countries. That is descriptive enough, or so I would have thought; my rough model of the London Whale position was a combination of basically long IG index synthetic credit by selling protection on 10-year CDX.NA.IG.9, untranched or senior tranches, and short higher-beta synthetic credit bits by buying protection on high-yield indices, junior tranches, something like that.

But it's also possible that the London Whale position is basically a blob of green glowing radioactive material that just deals indiscriminate pain everywhere it goes. So, for instance, this quarter, after causing massive and time-traveling losses last quarter and being mostly unwound and/or moved from the Chief Investment Office to the investment bank, it still managed to lose money in not one but two places - the investment bank, where the bulk of its ominously pulsing self "experienced a modest loss,"1 but also the CIO, where its mangled remnants lost $449 million on about a $12bn notional remaining position, or about 3.75% of quarter-initial notional.2

You can think a range of cynical things here. The most supportable, perhaps, is that CIO's daily VaR was $54mm last quarter3, meaning that the CIO's loss this quarter was a little over 8x its daily VaR, which is, um, high? A quarter is 65ish trading days; if you assume VaR goes with the square root of time then CIO's quarterly 95%-confidence-interval VaR was about, oh call it $449 million, meaning roughly that 95% of the time it would have lost less than it did, yet here we are. Of course there's the other 5% of the time, where the whale seems to live, but ... I mean, that is an odd number and might make you quietly ponder JPMorgan's new VaR model.4 BONUS FOOTNOTE!5

But the main thing to notice is this:

That is: the thing that the Whale sort of publicly "lost money on," that is, selling protection on investment grade credit indices, made money this quarter. And yet the Whale ... lost money again.

Trying to understand exactly what is up with the whalefolio will drive you mad, so don't, though you couldstarthere. But the simple directional view would be:
(1) Once there was a whale who bought a lot of credit protection to hedge his bank's credit risk.
(2) Later he went crazy and sold a lot more credit protection to hedge that hedge, somehow accidentally ending up massively long credit.
(3) Then he lost money, in part on credit moving against him and likely in larger part on second-order stuff - convexity, liquidity, dolphins - moving against him.
(4) Then he got rid of all that long credit stuff and ended back where he started, owning a lot of credit protection to hedge his bank's credit risk.

That, y'know, makes sense, because the whale's original job was to hedge JPMorgan's credit risk. And one thing that happens when you hedge credit risk is that, in an environment where JPMorgan is breaking records and credit is tightening, you will be caressed by a few gentle pleasant headwinds, and one of those will be that your hedges lose money.6 And, if you are a gigantic too-big-to-whatever bank with huge exposure to corporate credit in a quarter that has been world-historically kind to corporate credit, you might say something like, "yes, that is great, losing half a yard on credit hedges now is just fine, because we want our hedge position to be there when the hard times come."

Except they're unwinding the position as fast as they can. So, yeah. We'll miss you Whale!

1.Page 3 of the earnings release. That loss was not quantified. The IB synthetic credit bulk wasn't quantified either, in terms of notional, though JPM has said that it's $30bn of risk-weighted assets.

2.Page 13 of the earnings presentation has the loss; note (f) on page 10 of this quarter's supplement has the initial notional. Though those remnants were also "effectively closed out this quarter" (again p. 13 of the presentation), so I guess current notional is ~0 and the loss is like 7.5% of average notional? That's maybe a silly way to think.

3.95% confidence VaR, from page 42 of the supp.

4.Umm! I have little confidence in the mathematical fakery in that paragraph but I think it's baaaasically in the ballpark. Tell me if you disagree, but only in like a "VaR doesn't go with sqrt(T)" or "you are not accounting for risky drift which is material" way, not in a like "one-tailed vs. two-tailed" or "VaR includes weekends" way, I don't care about that. I suppose the most important objection is "CIO VaR" is, like, not just the synthetic credit portfolio, diversification blah blah blah, but, I dunno, still seems fishy. Note footnote (f) on page 42 of the supp, which says that their new model makes the portfolio VaR even lower, though in a way not totally clear to me.

5.Other, less supportable cynical thoughts might run to "why didn't you just write the Whale stuff down excessively last quarter, when you were taking a galumphing embarrassing write-off anyway, so you could cheerfully announce a gain this quarter?" Yourcynicism about that basic method of earnings management might be widely shared, but I guess you can't really mismark the portfolio that you're being criminally investigated for mismarking, even to mark it too far down.

6.Another is DVA.

Related