According to this study by the CFA Institute, three quarters of hedge fund managers think the outlook for computer-driven strategies is troubled. (HT Fintag).
The survey’s findings do reflect part of the long-held view that quant strategies have had their day. If pre-crunch those strategies were plagued by diminishing returns and a crowded marketplace, post-crunch they’ve been plagued by big losses, investor withdrawals and a few spectacular collapses.
By and large, it’s inadequate models that are blamed. But, as the CFA study’s authors write:
The conclusion of this discussion is that what appears to be model breakdown may, in reality, be nothing more than the inevitable fat-tailed behavior of model errors.
The models, in other words, are perfectly accurate, but inadequately applied.
Here, of course, it’s probably worth making a distinction: broken and inaccurate modelling has characterised a lot of the travails in the markets over recent months, but not always at the hedge funds.
Rating agencies in particular, were using some pretty outdated assumptions. In rating products such as leveraged super seniors, CDOs squared, or market value structures like CPPIs and CPDOs, rating agencies were using spread modelling techniques several years out of date; none of which had tails fat enough. Given how critical market liquidity was to all those structures, it becomes meaningless to talk of what happened as a “25-sigma event”.
Anyway, as the studies authors conclude, quant funds are most certainly not dead. The trick for them going forward will be to move beyond linear models based around Gaussian (or normal) distributions and instead trying to think of models which can move from “normal” market environments to “fat tail” environments dynamically and adjust investment strategies accordingly:
Eliminating the tails from noise would be an exceedingly difficult exercise. One would need a model that can predict the shift from a normal regime to a more risky regime in which noise can be fat tailed. Whether the necessary data are available is problematic.
Related links
Quants adapting to a Darwinian analysis - FTfm
Quando, quando, quando - for the uninitiated
Market Timing System using Grammatical Evolution - Alea
Predicting the credit crisis - FT Alphaville
