Sowood, So Long. And Not Short Enough.
Sowood, simply the next in the parade of likely credit hedge fund blow-ups. Exactly how many funds will be laid low by the current credit markets ugliness? I’d hazard a guess that the final count will be in the low hundreds. I keep asking myself why, WHY this is happening, and it always comes back to THE common thread separating the truly successful hedge funds with long track records from road-kill: strong risk management practices. This seems like straight-forward common sense, right? Well, apparently not. I know as much as all of you do, but what I can infer from Sowood’s situation is the following:
- They were long lower-rated credit instruments, either pure credit or credit spreads;
- The likely hedged the long positions by shorting equities and/or high-grade credit instruments;
- They likely employed leverage to amplify the bet; and
- Their book wasn’t very well diversified by strategy, even if it was across a range of credits.
The net result: lower grade credits eroded, higher grade credits eroded as well, and the drop in equities didn’t begin to offset the decline in credit values. And leverage only increased the adverse result of the trade. Now, consider that Sowood was a Day 1 $2 billion launch by Harvard Management rock-star Jeff Larsen. This is not some greenhorn tossing around huge institutional dollars without having any inkling of what they should be doing. This was JEFF LARSEN. Of the super-successful Harvard Endowment? Right - you know the one. Don’t you think Mr. Larsen knew better than to place so many concentrated and statistically-related bets such that, if all hell broke loose, he’d drop 50% of his NAV in a month? I’d think so. But then, I’d be wrong.
I guess if it can happen to LTCM and its brain trust it can happen to anyone. But didn’t Mr. Larsen learn from LTCM? Or, more recently, Amaranth? What is driving these types of behaviors? Unhappy with a few years of mediocre returns and trying to shoot the lights out? I’d bet a lot of money that Sowood’s true NAV didn’t drop 50% in a month - it actually dropped a lot less. Why? Because it really began dropping well before last month, it is simply that positions weren’t marked to true liquidation value but marked-to-model. I will almost guarantee you that this high-profile blow-up will cause many to revisit this issue - and fast. This kind of practice causes artificial stability in both position values and fund NAV, and generally provides a false picture of risk as well as possibly resulting in excess manager compensation. Autocorrelation - the smoothing of returns - is a big no-no, and if there is some basis in fact that banks aren’t causing gradual mark-downs in counterparty collateral because it would hurt their own proprietary positions, we’ve got a big, big problem. And this is what I’m afraid we may have.
Anyway, at times like these I hear the refrain from Queen “Ba da ba, ba, ba, another one bites the dust.” And I’m sure I’ll be hearing it a lot more in the coming months.
