Skip to main content

Trading Against Your Clients Works Best When You Lose Money

  • Author:
  • Updated:

This is sort of a strange footnote to the London Whale: one of the hedge funds that made money feasting off his carcass was run by JPMorgan*:

Even as a trader for JPMorgan in London was selling piles of insurance on corporate debt, figuring that the economy was on the upswing, a mutual fund elsewhere at the bank was taking the other side of the bet. ...

But perhaps one of the most surprising takers of the JPMorgan trade was a mutual fund run out of a completely different part of the bank. The bank’s Strategic Income Opportunities Fund, which holds about $13 billion in client money, owns about $380 million worth of insurance identical to the kind the “London whale” was selling, according to regulatory filings and people with knowledge of the trade. It is unclear how much the fund made.

This is ... not surprising. Some people want to sell CDS, some people want to buy it. That's how there's a market. And when you're as big and interconnected as JPMorgan, it's not surprising that the market often crosses between bits of yourself. That is, it's sort of silly to think of JPMorgan as a market participant; it is rather a nexus of many many market participants. Some of those participants are "JPMorgan," in that they're interested in the performance of JPMorgan as an entity; others of them are "clients" in the sense that they are buying securities from JPMorgan or having their assets managed by JPMorgan in some separate or mutual-fundy way; but to think of them all as JPMorgan is silly.

But the conclusions from this unremarkable fact are sort of interesting:

In that way, it is different from the example of Goldman Sachs in 2007, when it sold subprime mortgage securities while betting against them. In the case of JPMorgan, it was the reverse: the bank took risk with its own money to sell the insurance contracts that have cost the company money. The asset management division, meanwhile, invested on behalf of its clients when it bought the contracts.

In this case, it may turn out to be a silver lining. If nothing else, it indicates that the asset management division, run by Mary Callahan Erdoes, acted independently from the bank, as is required.

“You’ve got so many different businesses, they are not coordinated and they are not telling each other things and that turns out in this case to have been a virtue,” said Robert Litan, vice president for research and policy at the Ewing Marion Kauffman Foundation. “But that also feeds into another concern, and that is that JPMorgan is not only too big to fail but too big to manage.” ...

“There’s really effectively only one lesson other than the interesting irony,” said Douglas J. Elliott, a researcher at the Brookings Institution. “The lesson is that the asset management firms really do act like different bodies. They don’t share info. They don’t always have the view of the rest of the firm. That’s how we want it to be, and that’s how it was in this case.”

So, one: yes. I'm sure that's true - that Bruno Iksil and the managers of this mutual fund didn't sit down and talk out the risks and rewards of this trade and ultimately decide that one of them should go long and the other should go short.

Two, though: what if this had gone the other way? This trade is in some crude sense unnatural: you would naïvely expect JPMorgan's CIO to be buying corporate credit protection and its Strategic Whatsit Whosit to be selling.** Imagine it had gone that way: the CIO had bought zillions of dollars of index credit protection, the mutual fund had sold a chunk of it, and then ... well in this scenario imagine that CDX deteriorates, the CIO doesn't get picked off and makes $2 billion, and the mutual fund loses some money. Would the conclusion there be "the asset management firms really do act like different bodies, bodies housing brains that are sometimes smarter and sometimes dumber than the CIO," or would it be "ZOMG JPMorgan sold corporate credit to customers while betting against it"?

I don't know! The whole thing where they actually probably don't share info between asset management and CIO would probably be helpful. And JPMorgan making $2bn off a trade is probably less newsworthy than it losing $2bn. Still, I suspect the main reason that this trade "is different from the example of Goldman Sachs in 2007, when it sold subprime mortgage securities while betting against them" is that here JPMorgan-as-entity lost money and JPMorgan clients made money (maybe!).*** And that the one worse PR move for a bank than losing money for itself is making money trading against its customers.

As One JPMorgan Trader Sold Risky Contracts, Another One Bought Them [DealBook]

* Also not a hedge fund, whatever.

** That mutual fund is a credit fund; its natural state of being is to be long corporate credit though I suppose its purchases of CDX protection were a market hedge to being long specific names. (Or perhaps it is actually not that into corporate credit, whatever.) Similarly, the CIO is at least in part in the business of hedging JPMorgan's broader credit exposures as a bank, so you'd think it would be interested in buying protection (as it was in addition to selling).

*** I mean, there are also other, more sensible differences, like that GS's Abacus trades were marketed by the people who stood to gain on them rather than through independent brokers (bad fact for GS), or the fact that JPM's Structured Whatsit investment decision was made by JPM rather than its clients (bad fact for JPM if it had lost money).


You Say "Voldemort" Like That's A Bad Thing

Do you think that Bruno Iksil, when he woke up in Paris on Friday looking forward to trading from home in his black jeans, expected to become an international celebrity? The evidence suggests not. You may remember Iksil - possibly under other names like "Voldemort" or "the London Whale™" as the JPMorgan chief investment office trader who has sold protection on $100bn of notional of a CDX investment grade index to ... hedge ... JPMorgan's massive short position in credit ... or ... something?* Anyway a lot of people are mad at him because that's just too much protection to sell on that index and so they are complaining to Bloomberg and the Journal about how he is manipulating the market and also taking huge proprietary risks with JPMorgan capital that should obvs be regulated out of existence. This is weird in a lot of ways but one of them is that you can distill a lot of the Volcker-Rule complaints into "my God, you're telling me that JPMorgan is exposed to $100bn of credit risk on investment-grade debt issued by a diverse mix of 121 U.S. companies!?" No! JPMorgan is exposed to something like $750bn of credit risk on debt issued by a diverse mix of companies. Some of it's non-US. Some of it's not even investment grade. And that's just in its loan book.** Is writing $100bn of protection on the CDX.IG.NA.9 a terrible risk to take with investor and depositor and government-backstop money? Well, define "terrible risk." It's certainly less risky than operating the rest of JPMorgan.***