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.
——-
——-
COMMENT:
AUTHOR: Alex
EMAIL: alex@pinemountaincapital.com
URL:
DATE: 07/31/2007 07:03:27 PM
Rumor has it that Sowood was long secured debt, short subordinated paper (let’s say $4 short sub for every $10 long secured) as well as long equity puts (which didn’t help much). The problem, supposedly, is that the secured debt widened out MORE than the subordinated as banks unwound and now the senior spreads trade almost on top of the sub spreads, i.e. there’s no longer a discount for subordination! Obviously an attractive opportunity for the likes of Citadel. The whole affair is just another object lesson in how bets on statistical quantities that are not immediately arbitrageable can be run over by supply/demand imbalances. It’s qualitatively no different than LTCM, Amaranth, or the big interest rate option losses many I-banks suffered in the 90s.
——-
COMMENT:
AUTHOR: Soren
EMAIL: soren@dopeness.org
DATE: 07/31/2007 09:54:06 PM
I’m detecting a pattern here:
illiquid assests + leverage = bad.
——-
COMMENT:
AUTHOR: Bill aka NO DooDahs!
EMAIL: nodoodahs@aol.com
URL: http://billakanodoodahs.com
DATE: 08/01/2007 09:36:01 AM
If Alex’s rumor is correct, this is just another example of the “spread play” f*&^ing another hedgie.
LTCM, Amaranth, the list of “kills” owing to playing with spreads on leverage is just growing.
To me (amatuer observer) the greatest enemy of hedge funds is the retarded idea that spreads are “less risky” than directional bets.
——-
COMMENT:
AUTHOR: Tom
EMAIL: anon@anon.org
URL:
DATE: 08/01/2007 09:58:51 AM
How would you characterize contemporary strong risk management practices?
——-
COMMENT:
AUTHOR: Etoile
EMAIL: page@pagestarr.net
URL:
DATE: 08/01/2007 11:48:12 AM
Leveraging can be ok with illiquid assets (e.g. with term financing).
More problematic is getting a margin call when holding an illiquid position. Also problematic is relative value positions in illiquid markets- the p.b. can mark both your longs & your shorts against you.
——-
COMMENT:
AUTHOR: Jeff
EMAIL: jbramel@oakhillinvestments.com
URL:
DATE: 08/01/2007 11:54:59 AM
Alex may be right, but I like how Soren gets right to the point. Nearly every dollar lost in hedge fund blowups has been the result of too much leverage and too little liquidity. (And we can include the outright frauds in this analysis— the total dollars lost in hedge fund frauds, while great news fodder, are mere rounding error alongside the $10 billion-plus lost in “legitimate” blowups.)
But does this mean we should abandon high leverage? Not necessarily. It has serious hazards, but can also bring serious rewards. Just because driving at high speed is dangerous doesn’t mean we shouldn’t ever do it— we just have to be very careful about how we do it. No one wants to do their daily commute at five miles per hour, even though it’s decidedly safer than driving at seventy. But going seventy requires the presence of handling and safety equipment that reduce the risk of doing so: I would do it in a late-model sports car but not on a souped-up farm tractor. Similarly, when you’re traveling on the rutted dirt road of illiquid assets, you’d better slow down no matter what you’re driving.
I plead ignorance of Sowood’s specific circumstances, but I’d bet the car wasn’t tuned for high-speed off-road driving. Even if they had a risk management framework in place, I suspect they were relying on (a) some combination of backward-looking historical analysis, and (b) assumptions of a “normal” mean-variance framework, in their risk analysis. Guess what: In a world of leverage, squeezes, asymmetric distributions, and fat tails, these assumptions don’t cut it, and may be worse than no risk management at all if they lull an investor into a false sense of security. The heavily-loaded family minivan, traveling at high speed down a winding mountain road, hits a pothole, breaks an axle, and tragically plummets off a cliff.
There is no perfect risk management framework, but with effort it’s possible to construct something much better than traditional mean-variance VaR models— and doing so is critical if you’re playing with large amounts of leverage. Make sure you’ve got a well-tuned rally car before racing in the dirt.
——-
COMMENT:
AUTHOR: richmanpoo_r
EMAIL: richmanpoo_r@yahoo.com
URL:
DATE: 08/01/2007 02:09:15 PM
So how would a company like Citadel profit from taking over such a portfolio ? Given the uncertain outlook and indications of further credit deterioration / liquidity dry up, I would think that taking on this risk would not be acceptable even to a giant like Citadel. Maybe they have offsetting positions and are just taking profits ?
——-
COMMENT:
AUTHOR: cfageek
EMAIL: cfageek@yahoo.com
URL:
DATE: 08/01/2007 04:01:02 PM
The reason is Citadel probably has a stronger capital base and + wants to bear the risk I wasn’t surprised when i heard Citadel buying Sowood’s portfolio. It’s amazing to see the history repeating itself. B/c The exact same thing happened last fall as Amaranth collapsed. Citadel and JP Morgan bought Amaranth assets after the positions lost some billion dollars. Here’s the article
http://online.wsj.com/article/SB115877354146769064.html
Actually, I’d love to buy these illiquid assets from Sowood at deep discount. who wouldn’t?
——-
COMMENT:
AUTHOR: Soren
EMAIL: soren@dopeness.org
DATE: 08/01/2007 04:01:36 PM
@Jeff,
I wasn’t implying that leverage is always a bad thing. I DO think that a very simplified way to explain the problem (not universal by any means) is to say that the more illiquid the asset, the more risky using leverage becomes.
Take spot foreign exchange for example. Leverage in and of itself presents less of a risk because liquidity is extremely high and NAV can be marked to market in real-time.
——-
COMMENT:
AUTHOR: Yaser Anwar
EMAIL: yaser@yaseranwar.com
URL: http://www.yaseranwar.com
DATE: 08/01/2007 10:54:56 PM
It’s interesting reading everybody’s comments, but IMO its not about leverage or risk management, as much as its about portfolio strategy, more specifically sector-concentration and knowing when to fold. When you’re investing in illiquid markets, especially through complex trading strategies, you’re cognizant of the liquidity issues that will arise.
Most hedge funds these days use multi-asset class trading and risk platforms (no more one-size fits all ASPs, its all in-house servers, greeks, pre-configured ASPs) that add in support for trade processing; there are point solutions that address only specific types of securities, and most hedge funds use multiple prime brokers who look at the fund’s risk levels.
The very high leverage ratios of banks/financial institutions, such as (to name a few): ABK’s leverage is 80.8 to 1 (Face Value Bonds/Statutory Capital) and MBIís Credit exposure is $635 Billion based upon $6.8 Billion in capital vs. Citigroup where Credit exposure is $1.11 Billion based upon $127 Billion of capital; impeding housing bubble thanks to erosion of credit standards; and growth in hedge funds running, in this case sub-prime/ABX portfolios are a recipe for disaster.
As we know, for these illiquid strategies there is no mark-to-market price. This leads to hedge funds adjusting their performance data to minimize variations -) This leads to a bias that results from subjective decisions about the value of market positions at given times. And when you’re managing billions of dollars in illiquid markets, leading to fund concentration, alongside the other hedge funds, total value of sub-prime related derivatives was I think $240 billion, this leads to very high bias ratios (something investors should have been monitoring).
In such an environment it is only prudent to liquidate at least 6-7 months in advance. While it is easy to say all this ex-post, what surprises me is, given that Jeff Larsen was an emerging markets guy, his fund failed to realize the consequences of such concentration, at a time when Jacobe & Mobe funds were starting to trade these markets.
I’m not surprised that the secured debt fell more than unsecured debt. When you’re facing margin calls or in general the market starts to reprice risk premia, the unsecured debt, which is more illiquid, is harder to off load than secured debt (when funds look to reduce risk, they will look to sell the more liquid debt when markets are heading south, so you get 100s of them selling the supposedly more liquid debt, leading to higher supply than demand on a comps basis with the perceived to be higher risk unsecured debt which couldn’t be off loaded so easily.)
There is an increasing trend of positive correlation among previously uncorrelated markets. So, inspite of what a fund may perceive as a RV trade, the risk is not truly hedged.
——-
COMMENT:
AUTHOR: Eclectic Contrarian
EMAIL: econtrarian999@gmail.com
URL:
DATE: 08/02/2007 04:33:00 PM
A lot of things are unsurprising ex post.
I guess the only thing I’d add, which I’m sure many have heard before, is it can also be highly dangerous to assume historical liquidity when thinking about leverage alongside the return distribution.
Funny little personal observation. On the flip side to these terror stories is the fact that at least with smaller equities, share-based liquidity can actually get a hell of a lot better as their fundamentals (and stock price) improve. Helps with bets in those cool binary scenarios — in this case, sizing based on historical liquidity could actually under-size the optimal position.
——-
COMMENT:
AUTHOR: floyddbarber
EMAIL: floyd_d_barber1@yahoo.com
URL:
DATE: 08/03/2007 08:51:21 AM
Dealers are not forestalling a gradual markdown in collateral pricing. They are engaging in immediate and draconian price reductions.
——-
COMMENT:
AUTHOR: Michael White
EMAIL: tokintrader@mac.com
DATE: 08/10/2007 09:59:15 PM
Re: “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.”
Roger, certainly RIEF is the truest of truly?
From RIEF: ”August (down 8.7% through today[Friday 08/10]) is a different story. The culprit is not the Basic System but our predictive overlay.”
“While much of the damage was due to weak markets, our system experienced meaningful relative losses during the first two weeks of the month,”
Volatile World we live on right now. Quants might want to begin building ‘organic overlays’ - it takes awhile.
Too be continued…
——-
COMMENT:
AUTHOR: Brian Nelson
EMAIL: vikingace@yahoo.com
URL:
DATE: 08/18/2007 08:43:59 AM
It is truly amazing to this novice that hedgefund speculations{bets} are treated as if success is guaranteed. In this elitist banker system of course only millionaires and above qualify for this special treatment of using highly doubtful assets and then leveraging them to the moon and in effect jeopardizing billions of peoples’ next meals. When similar strategies were used in mortgage markets they were known as “liar’s loans” evidently when one has enough money an agent can be found to secure these “liar’s loans”.
——-
————
3 years ago
| view comments
| Hedge Funds
