I’m Pretty Sure JP Morgan Lost $2 Billion Just To Spite Me
May 11, 2012 27 Comments
So by now you may have heard that in (what I think is, though some may disagree) a delightfully entertaining turn of events, JP Morgan appears to have up and lost something like $2 billion on its whatever-the-hell trades. (Here’s the obligatory Matt Levine/Dealbreaker post.) Now people, I know that seems like a lot, but really, it’s just a number. Who’s to say one number is any bigger than another, after all. A loss is a loss. Like just the other day I Redbox’d the awesome Haywire starring the kick ass Gina Carano without realizing it was also next up on my Netflix queue and in the mail anyway (!). That’s something like $1.30 down the drain right there! Talk about dumb. See? Your eyes glazed over because, numbers.
Now as you probably can’t help but fail to recall I took some interest in the (still weird) publicization of JPM CIO and its trades at the time it all made the news. Relatedly, although this may be harder to recall, I have had occasion to write contemptuously about The Volcker Rule on The Internet. Like, oh, a few hours prior to JP Morgan’s big announcement for example. And I was pretty much ready to declare victory, too. I mean no one’s ever really argued with me and that must mean I’m right right? Mission accomplished. You LOSE Volcker Rule. I win. 1 to nothing. But as Craig Pirrong astutely points out, this whole losing-$2-billion thing may put a damper on my side of the debate.
So naturally what you came here wondering is this: Does this prove me wrong? Do I have to eat my words? Should I be embarrassed and penitent? Should I eat my hat or a crow or some other thing akin to a hat/crow? Is the Volcker Rule a good idea after all? Those are perfectly fair questions. Let me check. [checking...checking...checking...]
Ah here we go. All right then, I’ve checked and the answer is:
No. JP’s CIO office losing $2 billion on credit index/index tranche trades actually just proves my point. The Volcker Rule is still a bad idea. So there you have it.
Oh hey? You’re still here? All right then I guess I’ll elaborate.
The first thing to understand about these trades is that – all along – JPM has classified them as hedges. Hedges are not ‘prop trades’ and there’s no sane way to classify them as such (or more precisely: there’s no way to clearly define, for a given trade, which part/how much of it is a Kosher, Volcker-Approved Hedge and which of the rest/residual is Unkosher, Volcker-Disapproved Prop). The second thing to understand about these trades is that they were risky and dumb and didn’t work as hedges that well. We know that now because golly they have lost $2 billion dollars and hoo boy is that a lotta clams. But JPM apparently did not know that before learning it via the tried-and-true risk/control method of Noticing Large Negative P&L Emails Coming To Your Blackberry Right Before Dinner And Ruining The Rest Of Your Evening.
What this establishes, I think, is that trades can be risky and dumb and lose money without being ‘prop trades’. A trade can serve a ‘legitimate’ purpose – like it’s intended as a hedge, to ‘legitimate’ activity – and yet still lose a lot of money. Now you and I and Paul Volcker all agree that all else equal we don’t want banks to lose a lot of money. So a ‘rule’ like, say, Don’t Put On Trades That End Up Losing Money might make sense (if you also had a time machine and could check the future), but a rule saying Don’t Do Prop Trades makes no sense as a way to prevent risk, because it simply does not ‘prevent banks from taking on risk’. Exhibit A: JP Morgan. Right now.
And that is why it proves my point.
Now I’ll field some questions. First up is one of the Tyler Durdens at Zero Hedge who is hung up on insisting that JPM CIO was simply a ‘prop desk’ and this is a ‘prop trade’. My answer: I don’t care about such a debate. Remember my position is not that this trade, or any trade, is or isn’t a ‘prop trade’. It’s that it doesn’t matter, because I don’t think the concept can be sharply defined in the first place. JP can say Is Not and Zero Hedge can say Is Too but there is no such thing as an objective way to resolve the matter. Meanwhile, risk can be defined, numerically. So if you say you want to control risk, fine, control risk, with objective metrics and limits. What does whether that risk came from a ‘prop trade’ have to do with anything? And even if I concede this was a ‘prop trade’, by what objective rule would Regulators ever identify/prevent it?
Matthew Yglesias wants to call this a “loophole” in the Volcker Rule. The loophole being that you can hide a prop trade by calling it a hedge and the Volcker Rule allows hedging. Now in a sense, this is kinda my point (that it’s hard to tell the difference between ‘hedge’ and ‘prop trade’ because there’s no dividing line). So this would be an interesting loophole to point out if Matthew Yglesias (unless he’s suggesting that the Volcker Rule should close this ‘loophole’ by banning hedging, which would be retarded even for him) could tell us exactly where one can draw the line between hedging and prop trading. But of course he can’t (except after the fact), and Regulators can’t, again illustrating why banning a thing called ‘prop trading’ is misplaced.
Again the thing to remember is that JP Morgan themselves thought this was a good hedge. The Zero Hedge/Yglesias/etc. position is that JP Morgan has been lying about them being hedges. But what does that even mean? To know that, you’d have to define what counts as ‘a hedge’ and what doesn’t. Which none of those people – let alone Regulators – possibly can. Besides, you can’t just ignore that part of what was discovered and announced today was that JP Morgan made a boo-boo in its ‘rithmetic:
“There were many errors, sloppiness and bad judgment,” Dimon said as the company’s stock fell in extended trading. “These were grievous mistakes, they were self-inflicted.”
People are making a big fuss about JP changing/correcting/restating its VaR number. Now on some level I think this is overkill, since my longstanding official position is that VaR (while in theory is a reasonable thing) in practice is a load of crap and the one thing we know about it is that it is always wrong, every day. Their old VaR was wrong and their new VaR is also wrong (just maybe less wrong, thus more truthy?) and I guarantee you this.
But the revision does illustrate a point I want to make: wrong risk causes bad hedges. For example suppose – as now seems likely – they were (grossly, not just slightly) miscalculating their own risk. That would mean they were looking at tables of numbers that told them their risk was X when it was really Y. And that, in turn, could have meant that the Trades They Thought They Needed For Hedges (to reduce their risk) actually would not and did not do so. From there it’s very easy to see how they could end up with a trade they thought was a ‘hedge’ but really could Lose $2 Billion Dollars. This is what happens when you miscalculate your risk. But that doesn’t make the resulting trade a ‘prop trade’ (even though clearly, it was ‘risky’); it just makes it a…well…a mistake.
So let’s say I agree with you that the gov’t should ban banks from putting on ‘prop trades that aren’t really hedges’. Great. Let’s do it, Regulators! Let’s go identify some. We’ll do a hard-target search of every gas station, residence, warehouse, farmhouse, henhouse, outhouse and doghouse, and when we find trades that are ‘prop’, we’ll…well, we’ll write a strongly-worded letter to that bank asking them to take it off. Fine. But how are we going to do that? We’re going to have to look at numbers. Lots of numbers. Giant listings of positions. Summaries of those positions’ risk. Then we’re going to have to look at the trade in question. (Every trade, I guess.) And we’re going to decide whether we think the trade Really Does Hedge those (other N-1?) positions, or Doesn’t.
Which is the exact same thing that JP Morgan was doing here, and screwed up royally, you see. And we as Regulators would have to use JP Morgan’s own position reports and risk numbers too. (Are we, Regulators, going to independently calculate our own deltas and gammas and correlation risk?) In other words we’re looking at the same exact information that JP Morgan was looking at, and using – and which was WRONG. Because part and parcel of JP Morgan’s announcement is that they screwed up the numbers. So why/how would Regulators even identify/stop this ‘prop trade’ using numbers that (we now know) were wrong to begin with? At best, they’d make the same mistake JP made: ‘Yup. The numbers show it’s a hedge.’
There’s a more subtle view one can take though and it’s this: perhaps guys on the level of, say, Bruno-the-trader knew/mentally thought of this position as a prop trade mostly designed to make money; but guys on the level of Jamie the CEO didn’t, and kinda believed the it’s-a-good-hedge story based on the (wrong) numbers they were shown. This? Is possible. VERY possible. Likely even. Remember in my hierarchy of who I’d want to query about the logic of the trade, the CEO was way down the list, not far above Paris Hilton. Also remember that I don’t think there’s a sharp distinction between ‘prop trade’ and (other trades) in the first place. Putting on trades-you-think-of-as-designed-to-make-money, while staying-within-the-boundaries-of-whatever-else-you’re-supposed-to-do, isn’t just something that prop traders do, and it isn’t just something that JPM CIO does. It’s something EVERYONE does, ‘market-makers’ included.
So let me describe a scenario. While I still don’t quite know (and wouldn’t say if I did) whether the actual trade they had on was effectively a 5y/10y IG9 flattener, or an equity/mezz (go short equity for the ‘legitimate hedge’ and long mezz to ‘pay for it’), or an IG/HY decompression trade (short high-yield/long investment grade), or some combination of these things; and I don’t know whether the purpose was a genuine hedge, a ‘VaR hedge’, a (similarly) regulatory-capital relief trade, a DVA hedge, or a full-on punt/prop trade designed only to make a buttload of money, I think it’s reasonably safe to assume the flavor/spirit of the trade was something like this:
The [risk/VaR/reg cap calculation/DVA?] we calculate comes out to be X. Now, if we did trade 1 (a short, presumably) by itself that’d reduce/trim X so that’s theoretically the sort of thing we’re supposed to do, according to our official mandate. But trade 1 (being a short/insurance) is expensive/unsexy and will just bleed out money, making our desk look non-profitable and giving us no bonuses. But look, if we also do trade 2 (a long presumably) along with it, that will help pay for trade 1, and because of a quirk in the (wrong) risk calc, the numbers even say this trade package would still reduce our ‘risk’. So trade 2-versus-1 is more efficient, and it’s also an intelligent trade economically, and meanwhile, look how much money it should make us in [the most likely scenarios], Jamie!
And Jamie, being (I know this is a rare property in a CEO) a guy who likes for his firm to make money, liked that story. And so they did the trade. Two problems. One, as mentioned above, if the risk calculation that told you X was wrong in the first place, then all bets are off. Two, they seemed to have, unbelievably, forgotten that no matter how ‘intelligent’ this sort of trade 2-versus-1 looks on paper, to maturity – the mark to market can kill you. Especially when your position is dominating the space and gets publicized by some hedge funds; i.e. if you truly are a ‘whale’ – then (as Pirrong points out) you’re screwed. As ‘Walter Kurtz’ shows, IG has widened on this news, presumably due to the market lightening up longs/loading up shorts (and/or just speculating…) in anticipation of JP having to unwind these trades. So for those keeping track of where we are in my abridged-script version of events, we’re right near the end:
BRUNO: Damn, now I suddenly gotta go short like 20 kajillion dollars worth of IG9. Where am I gonna find someone who'll sell me all that protection?
A new BLOOMBERG CHAT WINDOW pops up on BRUNO'S SCREEN. Camera zooms in. The window reads: "HEDGE FUND TRADER: hey bruno u randomly happen to need any ig9 protection right now by any chance cuz i could sell you some for only like twice what i paid for it?"
Meaning, all of you who were weeping for poor hedge funds because JP Morgan was sullying the purity of synthetic credit indices with their trades can now rejoice: those hedge funds are probably closing out their 4-, 5-month positions at a decent profit. The baptist-bootlegger gambit always pays off after all.
Matt Levine wrote an epic 3-page followup here.
Kid Dynamite has a related discussion here.
Re: my conjecture of this trade as a decompression/long-short trade, this post at What the TF? explains the possible thinking in more detail.
What $42 billion of HY short going to cost? About $2.5 to $3 billion per annum. What??? You got to be kidding me?
No, we can’t do that. It would be too big of a drag on revenues. The hedge would cost more than the business makes. Besides, this is just for disaster protection. What can we do to offset this?
Well we could sell some IG. It would give us income to cover the short.