So all the financial news today

[giandujakiss]

is about how JP Morgan lost $2 billion on a bet that corporate bonds would do well (with a potential other $1 billion in losses forthcoming, I guess?)

This bet has been apparently public, and has been scrutinized in the financial press, for nearly a month - so much so that hedge funds began buying up the other side of the bet, which may have been at least part of the reason JP Morgan lost so much money.

This has also restarted a debate about the Volcker Rule, which is a rule that is designed to limit a bank's proprietary trading (trading for its own profit) while allowing it to continue hedge trading (trading to offset the risks a bank has merely by making its loans) - with the difficulty being, it's often hard to tell the difference.

The situation is incredibly complicated although it's often really hilarious to read insiders' discussion of it. That said, I found this very long post to be the best explanation - in the sense that it explains the technicalities pretty well. I don't know if it's for everyone - it may be too technical for some, or too simple for others, but it worked well for me. Some points I'd explain first, though - once people started detecting this trade, the trader was nicknamed the Whale of London or Voldemort, but his identity was actually revealed as a guy named Bruno Iksil.

And, just to be clear on the terminology: to go "long" on something is to bet that it will do well, to go "short" is to bet it will do poorly, and a hedge - as I explained before - is supposed to be a trade that covers your ass so that you aren't overexposed to a particular risk. To buy "protection" is to enter into a hedge that protects you against the risk of default - it will pay you money if the asset loses value. To sell protection is to promise to pay out to someone if the asset loses value. Therefore, when you "buy protection" you are making a bet that the asset will fail, essentially like buying insurance - you're taking the "short" side of the bet, to "hedge" against the fact that you already took the long side somewhere else. When you "sell protection" you're selling insurance; you're making a bet that the asset will do well and you'll never have to pay out.

Also, understand that JP Morgan has not revealed the details of what went wrong, so a lot of the posts on the subject are informed speculation.

Comments

Werewhales of London

[auroramama]

Very interesting, especially for the evidence that financial firms have either failed to learn that you need more than one type of model, or successfully learned that everything they do is rewarded sooner or later. I'm with the person (at dealbreaker, I think) who says you need stupid (rigid) rules in place to limit the damage brilliant models can do. In fact, the stupid rules themselves can be considered as a second type of model. Unfortunately, no one is interested in limiting potential upside when they're not the ones affected by potential disaster.

One commenter complained that investors are the only ones who suffer when a firm makes mistakes. I don't know whether they forgot about employees who don't make most of their income from bonuses, or have simply written off employees as stakeholders altogether.

Re: Werewhales of London

[kore]

HA, great comment title!

Re: Werewhales of London

[giandujakiss]

Re: the models - I don't know if you read the things the OP links at the end of his post, but one of the Felix Salmon posts in the links makes that exact point, i.e., that "dumb" model rules that place certain hard limits are safer, but traders keep insisting on more flexibility and then you have - this.

[kore]

//tries to read article

Okay so now you’re JPMorgan and you’ve got about $375 billion worth of securities in the CIO – alongside some $700 billion of loans (page 87 of the 10-Q), of which $115bn are in your commercial bank (lending to businesses – page 28), $70bn are in your investment bank (lending to bigger businesses – page 16), $240bn are in retail (basically mortgages and stuff – page 18), and $187bn are in card services & auto (credit cards, car loans, student loans – page 25). So you have $185bn of corporate loans, plus whatever chunk of that $375n CIO position is corporate bonds, which seems to be about $60bn (page 92).

//brain asplodes

I just, I seriously can't even picture that amount of money. GLOBALIZATION YAYZ.

[giandujakiss]

Heh, yeah. And plenty of companies operate on a similar or larger scale.

[saraht]

I found this very long post to be the best explanation - in the sense that it explains the technicalities pretty well.

Ugh, don't encourage Matt L., he's such an industry shill. I wish he would shut up forever and Bess Levine could go back to making almost all the posts. At least she's funny.

[giandujakiss]

I actually appreciate his perspective. Let's just say he sounds a lot saner than most industry shills, generally not quite as extreme, and I like hearing that POV.

[saraht]

I appreciate him primarily as a wonderful example of a dumbed-down version of the thinking that got us all into trouble in the first place: "Based on a far-too-simple model of how the world works, I think things should work this way; therefore they do work this way; therefore anyone who tries to change anything is just fighting reality and is going to mess everything up."