Don't hate on the QUANTS.

  • Thread starter Thread starter iosakwe
  • Start date Start date
Joined
4/24/07
Messages
46
Points
16
The DOW was down 400 points today and subprime is to be blamed.. but seems like others are pointing fingers at the MATH. Here's an interesting read from Bloomberg.

-----------------------------------------------------------------------------------------------------------------------
Highbridge, Goldman `Quant' Hedge Funds Lose Money (Update1)
By Katherine Burton and Jenny Strasburg
Aug. 9 (Bloomberg) -- Hedge funds that use mathematical models to drive investment decisions, including those run by Highbridge Capital Management LLC, Goldman Sachs Group Inc. and Tykhe Capital LLC, are losing money in August.
Highbridge's $15 billion multistrategy fund fell 4 percent this month, the Tykhe Portfolios Ltd. Fund declined 7 percent in the first three trading days of August and Goldman Sachs's $9 billion Global Alpha fund dropped almost 12 percent in the two weeks that ended Aug. 3, according to investors. A second Goldman fund is selling positions in response to losses.
``If the conditions change, the models don't work as planned,'' said Luis Rodriguez, head of risk management for New York-based Manhattan Family Office LLC, which invests money on behalf of an undisclosed wealthy family.
Hedge funds' quantitative, or ``quant,'' models have been confounded by wider credit spreads stemming from losses in the subprime loan market. The difference in yields between the riskiest corporate bonds and U.S. Treasuries has expanded about 1.5 percentage points since the start of June. Volatility, as measured by the Chicago Board Options Exchange SPX Volatility Index, has climbed to an average of 22 in two weeks from 14 since June 2003.
Goldman Outflows
Goldman's Global Alpha losses may lead to more redemptions. Withdrawals for the fund's $6.2 billion offshore version totaled $394 million in the month ended June 30, according to an investor who declined to be identified. That was almost three times the $142 million in new money added.
The fund decreased 8 percent during the last full week of July and was down 16 percent from the beginning of January through Aug. 3. There is an Aug. 15 deadline for Global Alpha investors who want to redeem money on Sept. 30.
The $29 billion Renaissance Institutional Equities Fund, overseen by a team led by James Simons at East Setauket, New York-based Renaissance Technologies Corp., also has faltered this year. The fund was down 0.4 percent through Aug. 3, after having been up 5.8 percent a month earlier.
Tykhe Capital LLC's fund fell 9.6 percent this year through Aug. 3. The New York hedge-fund firm, which manages about $1.8 billion, is trimming holdings, the Wall Street Journal reported, citing an investor briefed by Tykhe executives.
Man Group Plc, the world's largest publicly listed hedge- fund manager, named Tykhe in May to manage proceeds from its planned U.S. share sale. The offering by the London-based firm hasn't been scheduled.
French Bank
Two Bear Stearns Cos. hedge funds filed for bankruptcy protection in the Cayman Islands two weeks ago because of subprime losses. The New York-based securities firm then blocked investors from withdrawing money from a third fund. BNP Paribas SA, France's largest bank, stopped redemptions from three investment funds because it couldn't ``fairly'' value the holdings.
Goldman Sachs, based in New York, has sold positions in its North American Equity Opportunities hedge fund because of declines, a person with knowledge of the matter said. The fund lost 15 percent of its value this year as of July 27. It had $463.5 million in assets, according to a July 18 regulatory filing by Goldman Sachs Asset Management.
Goldman spokeswoman Andrea Raphael declined to comment.
Highbridge, which manages $37 billion from New York, said its largest fund was investing in Asian, U.S. and European equities. The fund has risen 5 percent this year, according to a letter sent to investors. A publicly traded fund with $1.8 billion in assets that uses no leverage has fallen 5.3 percent so far in August.
Brooke Harlow, a spokeswoman for Highbridge, declined to comment.
`Opportunities'
Highbridge and other funds are blaming declines on other quantitative managers selling positions.
``While we cannot predict when this technical pressure will subside, we believe it will present significant opportunities which we are well positioned to capitalize on,'' according to a letter. Fund borrowings are at the lowest level in its 15-year history, the letter said.
Black Mesa Capital, a Santa Fe, New Mexico-based quant hedge-fund firm, said an undisclosed hedge fund or investment bank is liquidating trading portfolios, MarketWatch reported, basing its information on a letter sent to investors. Black Mesa's own fund has slumped 7.5 percent this month.
To contact the reporters on this story: Katherine Burton in New York at kburton@bloomberg.net ; Jenny Strasburg in New York at jstrasburg@bloomberg.net
 
God

No one is this world can predict future. One person can do that....God can!

People have an illusion think that Quant can predict the market. I believe it was wrong for long time but nobody really accept the reality. First of all, QuantFinance is a useful tool to provide some indication of risk factors in the market. It will help to forecast, not even econometric.

Secondly, even we know the market is going down the next day but by how many points?
Traders or retailers enter the market at different time frame. Each time frame creates different sentiment which can affect the stock or entire market sentiment. Again, it may give a correct buy/sell signal but may be not.

Thirdly, modeling solely relies on historical data. We may try to filter the data to get the best unbiased estimator to hedge the risk so we believe, this is a winning formula. Does this make sense? Can we estimate "Qualitative Factor". For example, market was up on Wednesday because of Cisco provided some positive feedback to the market. ONLY CISCO. The underlying market condition is very ugly. Sub-prime loan is undone business. There are more to come in the next few weeks......

Can quant predict? I think "information" is the key.

Market experience is very important too...lile Mr. Alan Greenpan (Sorry if I spell this Guru's name incorrectly).

Cheers,
Kel

I am back after a long while....I do battle with the market and continue to do so..!:dance:
 
Historical data play a big role in forecasting. Thats why we study regression, time series analysis and other subjects. For example, we all can predict that sale of item X will go up before Christmas :) We don't know exactly by how much, but can construct the confidence interval based on the sample. So in certain areas "one can predict the future".

In Quant Finance, things are a little bit more complicated. But who knows, maybe one day we will have some sort of Extra Multinomial Regression / Super GARCH / Cool Factor model that will enable us to predict market movements just like we predict Christmas sales :) Or maybe this model already exists.
 
God will tell

Historical data play a big role in forecasting. Thats why we study regression, time series analysis and other subjects. For example, we all can predict that sale of item X will go up before Christmas :) We don't know exactly by how much, but can construct the confidence interval based on the sample. So in certain areas "one can predict the future".

In Quant Finance, things are a little bit more complicated. But who knows, maybe one day we will have some sort of Extra Multinomial Regression / Super GARCH / Cool Factor model that will enable us to predict market movements just like we predict Christmas sales :) Or maybe this model already exists.

Ideally yes. Your models will only predict well when there is a trend. So, regression and time series come in handy. What about in random walk situation, can you?

In normal situation, we know that people will spend more during Christmas. Also, like Chinese New Year period. What about the Sub-prime loans condition persists and we get a recession. Do you think your prediction now is valid?

All our mathematical tools are good to have but market sentiment and psychology play crucial part. You can't estimate how much is the percentage of shareholders will approve the takeover, unless you are an insider. Again, you trade on insider information is illegal.

If the takeover news materialize usually the stock of being takeover will go up and the other go down. Can you predict shareholder approval percentage? What information can you integrate into your model? This is all artificial information.

You may be lucky to guess it right! If you really can predict, you should be richer than Warren Buffet by 3 years from now. =D>
 
Many are indeed losing on this sub-prime mortgage debacle. But this is not a one-sided deal. It is a zero sum game, and there are some out there winning big, making millions, because they made good bets and went the opposite direction.

I think the media is being a bit paranoid. The reason behind this histeria is that some firms big name firms, like bear stearns, Goldman Sachs, and now Renaissance technologies have serious exposure to sub-prime mortgage.

But overall things are not thad bad! Not all firms are engaged in MBS or other structured-credit products; those who have had minial exposure to those products are now rejoycing their decision of not doing so! I think the bigger issue is of ethical nature. The SEC has recently opened an inquiry to further understand how those big banks woed uninformed investors into signing up for those sub-prime deals.

An open question now! Do you think that wall street firms take on more aggressive bets as their profits increase? I am sure many, who engaged in sub-prime bet, had foreseen this a possible meltdown, but they probably outweighted the potential possible gain over the risks that could follow. Greed can give you plenty of sleepness nights!

Again this is my 2 cents. I don't have a lot experience in the field, so some of my observations can be wrong.
 
We cannot predict the random walk. And we cannot predict people's actions. Maybe we should invest more in behavioral finance.
But still, we can learn something from statistics :) thats the reason we use correlations and related things. I'm just saying the world is not entirely unpredictable.
With zero-sum games it becomes a 50/50 situation where some win (how? maybe they were able to predict?) and some lose (so their predictions failed).
 
I think the media is being a bit paranoid. The reason behind this histeria is that some firms big name firms, like bear stearns, Goldman Sachs, and now Renaissance technologies have serious exposure to sub-prime mortgage.

I am not at all clear that the Goldman and Rentec losses are on MBS, despite recent equity market turmiol being clearly linked to the mortgage market.

My two cents, Goldman's Global Alpha being down 30% in a week is as good of a reason for paranoia as any. Does anyone not remember LCTM??
 
My two cents, Goldman's Global Alpha being down 30% in a week is as good of a reason for paranoia as any. Does anyone not remember LCTM??

Agree. The fact that central banks worldwide are throwing money into the inter-bank lending market to keep it from freezing up completely is another clue. Much more than after 9/11.

To hell with inflation, I guess. They can worry about that again once this crisis is past.
 
Back
Top Bottom