Reuters had a neat article today about how JPMorgan's CIO embarrassment increased credit spreads for a bunch of investment grade companies. The 121 companies included in the CDX.IG.NA.9 index, in which JPMorgan apparently had a $100bn long position, saw their CDS spreads spike in the days after JPMorgan revealed its losses - and its intent to unwind that position - last month. As Reuters puts it, those companies' CDS spreads
became more like the pawns in a battle between JPMorgan and hedge funds on the other side of its bet. This struggle so dominated a corner of the market that it sent false negative signals about the credit quality of some major companies whose underlying finances were largely unchanged, market experts said.
JPMorgan, sort of strangely, disagrees:
A JPMorgan spokeswoman said there was no causal link between the credit derivatives prices and the trading tied to the bank's losses. The theory, she said in an emailed statement, "is wrong and ridiculous."
But the Reuters analysis showed the 121 companies underlying the index of credit derivatives at the heart of the trading battle had a sharper increase in default insurance costs than 41 companies in a separate index that was not believed to be part of the big bets.
That statistical analysis - CDS on companies in the index went up by more than CDS on some other companies - is more suggestive than compelling, but also more suggestive than "wrong and ridiculous." I like suggestive when I can pair it with a story. What is the story - the actual trade that would do this?
One of the early theories about the Whale's detection was that he was trading against hedge funds who were doing "skew trades," selling protection on the index components while buying protection on the index. Protection on the index (which the Whale was selling tons of) was cheaper than protection on all of its components, so they bought the former and sold the latter expecting the gap to close. When it didn't close, because the Whale kept selling more protection, they started complaining to the media.
On that theory the trading against the Whale would have artificially driven down the CDS prices of companies in the relevant index in the weeks before they started complaining, as hedge funds sold lots of protection to hedge their purchases of index protection from the Whale. In May, as the Whale started unwinding his trade, they started unwinding against him - selling back to him the index protection he'd sold them, while buying in the protection they'd sold on individual names and so pushing up single-name prices.
The story and the statistics check out so I'll buy it. But I won't overpay for it; in particular, this "collateral damage" isn't worth getting worked up about. Technicals - supply and demand for an instrument driven by something other than a deeply analyzed desire to own or avoid that instrument - are just a part of life. Sometimes they work for issuers - Facebook, for instance, wanted to be in the Nasdaq 100 index so badly that it convinced Nasdaq to waive its index requirements to let it in early; this is not a matter of cachet but one of access to $50bn of Nasdaq-100-indexed money. Sometimes they work against issuers - talk to, for instance, the wee retailers whose stocks get crushed when the XRT ETF rebalances.
Both effects presumably happened here. The Whale's hedge fund counterparts - though not him, he wasn't trading in the underlying CDS at all, or so it appears - would have driven down the CDS spreads of some companies earlier this year, and driven them back up last month. Maybe they increased volatility in those spreads - CDS levels were generally widening in May and tightening in December-March, so these technical effects were procyclical, increasing risk signals more when they were already increasing for "fundamental" reasons.
Perhaps worth worrying about, but hard to avoid: every security is a "pawn in the game" between people who want exposure to its risk, or some nth derivative of that risk, and people who want to avoid that exposure. So here JPMorgan's unwind arguably sent "false signals" about the risk of, say, McDonald's debt - but JPMorgan was unwinding a trade because it realized that it had excessive and dangerous exposure to the health of investment-grade corporate America. That realization of excessive exposure filtered down into signals about the creditworthiness of investment-grade corporate America, including McDonald's. It's not anything McDonald's did - but nor is it wholly unrelated to the perceived risks of McDonald's.
There is perhaps another reason not to overstate the case. This sure sounds bad:
Yet the cost to insure against a default at CSX surged 28 percent to $64,300 for five-year protection on $10 million in debt on May 14 from $50,100 on May 1. At the same time, the company's stock fell just 5 percent. ...
There was a similar pattern at McDonald's. The cost of protecting the fast-food company's debt against default rose more than 19 percent to $24,300 on May 14 from $20,400 on May 10, while McDonald's stock fell just 1.1 percent.
Senior claims (debt) on one company should be less volatile than junior claims (stock) on the same company, so it's weird that changes in CDS prices were a large multiple of changes in stock prices. Except, no - comparing yields (of CDS) and prices (of stocks) is misleading. Here is a rough recasting into comparable terms:
A move of four basis points in McDonald's CDS spreads sounds - well, tiny expressed in basis points, but large expressed as a percentage change ("a thing went from 20 to 24 - a 20% change in the thing!"). But that doesn't change the value of McDonald's debt all that much - in fact, it changes the value by about 0.2%, a fraction of the change in the value of its stock. CSX's stock was down by ten times as much as the decline in its debt value implied by CDS levels. Which suggests that the phantom worries of the CDS markets were small relative to the real worries of stock markets.