http://www.zerohedge.com/news/8bn-loss-or-was-jpmorgan-unhedged-long-and-wrong-post-ltro2

An $8bn Loss Or Was JPMorgan 'Unhedged, Long-And-Wrong' Post-LTRO2?

Submitted by Tyler Durden on 05/22/2012 23:53 -0400

The full set of DTCC data is in (that is the repository for reporting CDS data) and reading between the lines provides us with some significant color on what was occurring at JPM's CIO office. For the Cliff Notes' version - see the summary at the bottom...

First things first, the position does not appear to have any HY9 tranche involvement at all, but a modest short HY credit index position was unwound in mid-Feb (we suspect related to the IG9 tranche unwind - since the delta ratio makes some sense at 1x ($25bn HY) to 5x ($120bn IG tranche). But nothing has been done since then.

However, when we look into IG9 gross and net notionals between tranched (the tail-risk hedge we believe they put on originally) and UnTranched (the index market where the whale managed to delta of the tail-risk hedge) the story becomes both confirming of our hunches and also very concerning.

The charts below tell the story of an early unwind of the Fed-induced-failure tail-risk hedge but an arrogant momentum chaser that left the massive long credit index position (hedge of a hedge) that had been the cause of dislocations in the Index-arb business (that other media entities have focused on) flapping into and after LTRO2 into around early-to-mid April (when we are sure Jamie got the call).

The drop in gross tranched notional was around $65bn in mid-Feb and a further $55bn this last week - which provides us with approximate sizing of around $120bn for the whales' exposure(tranched - in senior risk so modestly unlevered) - which is in line with the insider estimates of $100bn notionals. Furthermore the gross untranched (index) notionals dropped by around $100bn also which at aDV01 of ~$45mm implies the $2.5bn loss from the 50bp index shift alone.

Add to this the massive loss they would have incurred to get the remaining $55bn of tranche off the books in the last week (as a guesstimate - based on moves in the equity, mezz, and index positions that are the 'opposite' of the tail-risk hedge) around 10% of notional or a further $5.5bn loss) and the total loss is around $8bn.

However, there has been huge 'technical flow' in almost every liquid credit index (IG18, HY18, HYG, and JNK for example) which would have reduced the loss - though left considerable basis risk(hedging the loss with an imperfect hedge). Perhaps given the tranche unwind last week, the rally in credit indices this week reflects some more unwinds of the tail-risk's hedge and a slowing of the technicals in the market - leaving just weak fundamentals - though the basis risk (the difference in dynamics between the two side of the trade) could be very painful.

Think of the notional amounts as measuring the market's aggregate exposure to various indices - the details of net vs gross are important but do not change the narrative below...

1. HY9 net notionals are minimally changed the entire year - both tranched and untranched - the drop (and green arrow) is a tiny $0.6bn drop in net notionals (check right had scale).

2. HY9 Gross notionals show a drop of around $25bn in mid-Feb - which would fit with a beta-neutral HY hedge for an IG position - this was a short-cover in our opinion - which is a shame since it would have saved JPM a lot if they kept it on...

3. The main story is here!

In Mid-Feb, as with HY9 above, it looks like a decent chunk of the original tranche (tail-risk hedge) was removed (red line below). Whether this was in anticipation of LTRO2 - perhaps on the back of the market's massively strong momentum post LTRO1 (and his likely losses) - the whale appeared to decide to leave the index (blue line) long credit position in place (even though he had reduced his tail-risk hedge position dramatically.

This is the momentum-jockey arrogance of knowing better and having an unlimited capital account to draw on... in a nutshell - he had a hedged position into this point (that admittedly the delta was probably outperforming his original hedge quite well) and thinking he knew better, he just decided to get naked long credit (orange oval)...

Of course this basically marked the top in the market and instead of LTRO2 providing more ammo for the rally, it signaled the end of easing and market dumped. This put his hugely naked long credit underwater as he had less of a hedge against it now...

And as the chart below shows - the unwind (black vertical line) relieved some of the pressure on the IG9 skew (the spread between the index and the intrinsic value of the portfolio) - which had been stuck massively wide for months (orange horizontal arrow) until this unwind which allowed the arbs to get their payoff (dark red arrow)...

As is clear in early April - Jamie got the call and the position was cut aggressively(blue line)...

And this last week - we suspect the rest of the tranche (red line) was unwound (painfully)

4. As is clear below (red line) - this position was around 15% of the net notional of the marketplace - the IG9 market IS the tranched credit market and so this must have cost JPM a fortune to tear-up or persuade someone to take it on...

But - Index net notionals have not shifted much - suggesting the IG9 long (at least in part) is still on (or more likely hedged as we note below).

5. Here, by way of example is themezzanine tranche of the index (the 3-7% tranche) which jumped from 20% upfront to 31% upfront in the last few weeks. If we adjust for deltas and add in the equity and other less senior tranches to the Whale's tail-risk position (and adjust for the index shift) since being short one part of the capital structure is similar to being long everything else against an index short - we guesstimate the 10% of notional cost this last (or around $5.5bn) is not far off - especially given the illiquidity...

Which leaves an $8bn loss - though the last few weeks has seen indices all over the place grabbed as basis hedges so that the entire index position didn't flood the market. By way of example, IG18 notionals have soared recently...

So, in summary, it appears that the CDS data confirms what we suspected.

A large (~$120bn) tail-risk tranche credit hedge was placed.

The hedging of that hedge became very onerous but surprisingly profitable as markets rallied day after day with no give-back.

This led to a greedy trader lifting some of the original tranche (and the HY short side) and leaving himself much more naked long to the market into LTRO2 - which marked the top. Losses escalated through April (~$2.5bn or so).

Dimon went public (with some of the details).

Last week, the rest of the tranche was dumped (we suspect) at a large cost (perhaps ~$5.5bn) leaving, we suspect...

A potential ~$8bn loss and a heavy IG9 long credit position hedged (with major basis risk - difference in dynamics between the legs of the trade and the hedge) by various other liquid positions including shorts in HYG, JNK, IG18, and HY18 (and we would suspect equity/financials too).