The Volokh Conspiracy

Hedge Funds Are Beginning to Disclose Recent Losses.--

Among the many hedge funds disclosing recent losses is the main one run by Renaissance Technologies. Jim Simons, the spectacularly successful hedge fund manager of Renaissance, just reported that his main hedge fund has lost 8.7% so far in August. Nonetheless, in a letter to investors, he claims that "The culprit is not our Basic System" of investing (no online link yet). (According to Wikipedia, Simons’s personal compensation was $1.7 billion in 2006 and $1.5 billion in 2005.)

Hedge funds that severely restrict withdrawals may not be subject to severe challenges in the next few weeks, but others that don't routinely restrict substantial redemptions from investors may be selling parts of their portfolios — both to deleverage their investments and to raise cash to pay investors.

For the last few months, I have been developing and forward testing some statistical models to predict daily moves in the US stock market, ETFs, and some no-load, no-fee mutual funds. I noticed that in late July, my models ceased predicting moves in both the stock market and in commodities. I then completely redid the models, which have performed well in August, but they are much less consistently correct than they were in the May to mid-July period. Things have changed from the way they were from mid-2003 to mid-2007.

DDS:
You mean your redone models have performed well in the seven trading days we've had so far in August? Is that useful data?
8.10.2007 1:47pm
Dan Weber (www):
For the last few months, I have been developing and forward testing some statistical models to predict daily moves in the US stock market, ETFs, and some no-load, no-fee mutual funds


Oh, man. . .
8.10.2007 2:11pm
steve (mail) (www):
It's the new institutional ($29B) fund that is down. As I understand the old Medallion fund (which has posted one of the best records in the last 20 years, and is now closed to new investors -- most of the money in it is Simons' and employees') is still up.
8.10.2007 2:12pm
Lloyd George:
Anybody can make financial models and enjoy success in the short term, but they can be imperfect as everyone is now discovering. One critical flaw is that the models frequently don't take into account other entities using the same models. Further, the models don't take into account that, periodically, old assumptions get overturned and the human factor comes into play. So, it all becomes, garbage in, garbage out.
8.10.2007 2:14pm
Lloyd George:
Oh, I am enjoying the sweet smell of financial disaster in light of the latest complaints for associate attorney quests for pay raises. Kind of hard to enjoy your pay raise when you get laid off in the subsequent economic downturn.
8.10.2007 2:15pm
James Lindgren (mail):
DDS wrote:


You mean your redone models have performed well in the seven trading days we've had so far in August? Is that useful data?


Perhaps not.

Just to be clear, it's very easy to develop and back-test models that would have performed wonderfully in the past. The test of any model is its ability to predict the future, which is extremely difficult, if not impossible.

So last spring I developed and back-tested models. Then, using these models, almost every evening I made daily predictions of the performance of the market and various funds the next day. These models worked well until late July, when they performed very badly. I then redid the back-tested models including late July data.

On balance, those new models have done well in August in real time (predicting the next day's performance of the market and of particular funds). Nonetheless, the August returns are so variable that I have little confidence in these models, so I have committed only a very small amount in an old IRA account to using these new models to trade.

I hope that answers your questions.
8.10.2007 2:53pm
Shivering Timbers (mail) (www):
Just to quibble a bit with this statement:


Hedge funds that severely restrict withdrawals may not be subject to severe challenges in the next few weeks


Even hedge funds without any withdrawals may be subject to sizable margin calls if they're highly leveraged. One of the drivers of the current downturn is banks marking-to-market these illiquid mortgage-backed securities, which is eating into the margin cushions and forcing some funds to either sell into a dropping market (bad) or raise new equity (impossible).
8.10.2007 2:54pm
anonymous coward:
Yeah, when Medallion's down it'll be news.

The mean hedge fund has done okay recently, it's just that some have had great months and others have blown up.
8.10.2007 3:24pm
dearieme:
The test of any model is its ability to predict the future, which is extremely difficult, if not impossible.

Except for climate models, obviously.
8.10.2007 4:02pm
James Lindgren (mail):
Shivering Timbers,

Good point.
8.10.2007 4:16pm
Smokey:
Except for climate models, obviously.
Yes, that goes without saying, since climate models are infallible.

/s
8.10.2007 7:11pm