Now, the way quant hedge funds have operated is that when things went wrong for one firm, the odds were that its effected would be diluted by better performance from others. Usually, the odds of a lot of firms being affected by the same problem are rather low and consequently, their effects on the economy would be limited as well.
Ideally, I would like to say extremely rare. But after August and after the way the market has been looking lately (shaky) in the hedge-fund world, things are beginning to look a lot less uncertain. For instance, Paul Kedrosky talks about how a lot of quant funds seem to bleeding away, with steady losses almost on a daily basis.
So, going back to HFN, they make a rather interesting observation in their article, Twilight of the Quants:
Goldman Sachs dubbed the rapid drop in its Global Alpha fund a “25-standard deviation event,” meaning it would normally occur once every 100,000 years.
Of course, with Long Term Capital Management, virtually the same statement was made about their model.
David Nissenbaum, a partner at the law firm of Schulte Roth & Zabel, says he spent a lot of time this summer on the phone with hedge fund clients. There were margin calls, issues with illiquidity, you name it. This was a historic event that was happening.
The problem is, these “historic events” are happening every three to five years.
Of course, the one interesting thing is that there is probably a simpler explanation that could be at work here (i.e. if some hedge-funds are selling short, this triggers a cascade effect). And post August ‘07, who could blame them?
I mean, just look at this graph of the Required Leverage Ratio percentages over the years (from Paul Kedrosky’s chart) -
That does not particularly look good, does it? Combine that with the fact that even post-August, the average hedge-fund seems to be performing below par.
Now, the one other thing is the fact that while a lot of these early-adopter hedge-funds had significant leverage in the beginning, almost everyone today uses some sort of algorithmic methodology to quantitatively model higher level derivatives. So, the first-mover advantage is probably running out.
That begs yet another question — does this mean that all these models lead to Rome, or that we are just looking at a shake-up of sorts? The kind that happened post-DotCom-bubble?
« Hide
RSS 2.0
Add to Technorati
Save to Del.icio.us
Take it back