Wednesday, September 21, 2011
Credit Suisse Arb Index Fishy
Credit Suisse has an 'Arbitrage US Index' (CSIARBUS Index), that presumably reflects the return on 10 stat arb strategies. The total return since inception in Jan 2002 is above. It's a Sharpe of about 2.4, Madoff-like. If it's real, everyone should put most of their money in such strategies asap.
I'm rather skeptical of such indices. Where's the infamous August 2007 draw down? The index was first made available on Bloomberg in February 2011, and just this August this index had its worst draw-down ever. I have a feeling it is filled with innumerable back-fill biases, a typical naive backtest quants present all the time, oblivious to their selection biases. I looked on the internet for some kind of description, and found nothing.
The world doesn't need more institutions presenting implausible returns. It sure doesn't help their credibility.
Subscribe to:
Post Comments (Atom)
7 comments:
SGIXDVU2 INDEX is another example
basic question - if it is dodgy, how did they get away with it for so long? it is difficult to assume plain fraud, like madoff, or stuff like fake trades with non-existent cpties. control functions can easily check the valuation of the assets behind, it's "US names". it's simple arithmetic, not fancy mark to model.
As you said, the index got launched this year and the rest is just the "theoretical" performance if it had been run over the last decade i.e. just backtesting, not real investable returns. The chart of bbg gives the wrong impression to naive investors that the index has performed in the past which is far from the truth. Anyone can construct a index that has performed brilliantly over the past time series, especially when I have 10 strategies to combine!
that clears my question. all bb says is "begins 01/01/02". I guess they should show distinctly the part that is just hypothetical.
Check out SGIXEM6, Soc Gen's "Emerald". In my experience, all of the complex structured index products the investment banks create unrealistically optimised and then peddled oblivious buyers. It's an easy game to estimate the issue date of the structured product on one of these indices based only on the index line (eg eyeball when the sharpe go from >2 to <0)
All references to CSIARBUS on the Web seems to be to Eric's post or echos thereof. Also I can't find an actual traded product based on the index, though CS does offer similar ETNs, CSMN based on index HSGMN, and CSLS based on index CSLABLN. Both those indices have authentic-looking historical data, unlike CSIARBUS.
Absolutely, you have nailed it Mr. Falken. The hindsight bias stuffed index is one of the greatest (and costliest) investment hoaxes of all time. Line up the suckers!!
Post a Comment