Hedge Funds and Risk Software

NickL · Post by **NickL** » Thu Aug 17, 2006 12:33 am

@Nonius: That sounds interesting, what's the cost of maintaining it, I mean how much effort is it?

Do you also enter e.g. unconfirmed news or messages?

@Eric: Such a data repository is extremely valuabel, for sure, while vendor solutions have in common that they don't work but are expensive.

Nonius · Post by **Nonius** » Thu Aug 17, 2006 7:34 am

yes, we monitor rumors and general reputation issues, whether proven or not. unfortunately, unlike law, in banking (and business in general) there is no concept of innocence until proven guilty and there certainly is nothing like proof beyond resonable doubt. bad reputation spreads like cancer, whether warranted or not.

monkeyA · Post by **monkeyA** » Wed Aug 23, 2006 11:27 pm

Nonius, so do hedge funds (including the most sophisticated) use any kind of risk capital measures? Or do they just look at VaR and try and keep that to a certain level?

Do they use any risk based criteria when entering into a trade, e.g. effect on portfolio VaR, expected risk adjusted return on capital, etc.?

aaron · Post by **aaron** » Thu Aug 24, 2006 12:47 am

<< yes, we monitor rumors and general reputation issues, whether proven or not. unfortunately, unlike law, in banking (and business in general) there is no concept of innocence until proven guilty and there certainly is nothing like proof beyond resonable doubt. bad reputation spreads like cancer, whether warranted or not. >>

Unfortunately? You think it would be a better world if I had to go before a judge and win a criminal conviction before I could take my money out of a hedge fund? Or avoid putting it in in the first place?

Hedge funds organize themselves to avoid most disclosure and regulatory requirements, which makes them more efficient and flexible, but also removes some investor protections. You can't do that and complain that people are now judging you based on your reputation, instead of the financials you don't publish or the regulations you don't adhere to.

Nonius · Post by **Nonius** » Thu Aug 24, 2006 11:07 am

Monkey, a number of funds use a number of different types of risk capital measures. in some cases, it is VaR. in other cases, each trade is assigned a stressed worst case loss and that is the capital.

Aaron, it is unfortunate because rumors are, well, rumors. I visited a fund that had 30% of its capital pulled out by investors because of some rumors published in a few dodgy papers. In fact, the rumors turned out to be untrue. I'm not saying that one goes before a judge, but that one should exercise good judgement in deciding whether to invest or divest. This would mean that one wouldn't pull money out at the slightest instance of a bad rumor.

hedge funds avoid disclosure to *regulators*; it is the investor's job to probe as deeply as possible when performing due diligence. financial statements are always given to investors, so, not sure what you are talking about. as for regulation, I think bankers and investors are imposing a fairly large amount of de facto regulation. we sure do.

monkeyA · Post by **monkeyA** » Thu Aug 24, 2006 11:05 pm

Thanks Nonius, are there any public reference to the types of shock testing they do? do they add up the results of each shock, or correlate them in any way?

I am interested in how it compares to pension fund and insurance company risk measures

Nonius · Post by **Nonius** » Fri Aug 25, 2006 11:05 am

I don't know of public references. My personal experience is that if they take the stress approach, then it is additive. That is to say, compute stressed capital on a "trade", which may include hedges, and then add up all the stressed capital amounts. for a lot of strategies, it is very dangerous in my opinion to assume correlation. correlation, as you know, can break down at the worst time. take merger arb. hard to model using var. so, you try to estimate what happens if a deal breaks (assuming you aren't chinese direction). you could sort of think you have some diversification in a portfolio of merger pairs...then, out of nowhere, a market crash, which blows out a lot of the pair strategies. anyway....

Johnny · Post by **Johnny** » Fri Aug 25, 2006 1:05 pm

For systematic trading, clearly it's possible to run the whole portfolio + systematic responses through any scenario, including historical scenarios such as Oct-87, Aug-98, blah blah. For discretionary trading you can run the portfolio through these scenarios, but obviously this doesn't include the effects of changes to positions through the scenarios.

Nonius · Post by **Nonius** » Fri Aug 25, 2006 1:09 pm

true, they ignore the action taken upon a scenario happening. nevertheless, this is what many discretionary funds do.

Johnny · Post by **Johnny** » Fri Aug 25, 2006 1:13 pm

Sure. And actually, that's a very reasonable conservative thing to do. And it's consistent with the assumption that the shit hits the fan before you have a chance to do anything about it.