Predicting alpha: Not that hard after all finds new study
Jan 22nd, 2009 | Filed under: Performance, Analytics & Metrics, Today's Post
Regardless of their underlying investment strategy, it’s reasonable to expect that all funds should be comparable at some common level. The proof is in the pudding. Or, as we like to say, it’s not about “hedge funds” or “mutual funds”, it’s all about alpha.
While the calculus for comparing hedge funds and mutual funds on an “apples to apples” basis seems to be slow in coming, a gaggle of academic studies over the past 2 decades have proposed performance measures that compare the value-added by mutual fund managers regardless of the underlying style.
One such metric – called “Active Share” – was proposed in 2008 by Martijn Cremers and Antti Petajisto of Yale (see previous post). This metric is based on the divergence of the fund’s holdings to those of a passive index. Cremers and Petajisto found that funds with a higher Active Share tended to perform better than those with a low Active Share.
The notion of Active Share is intuitive, but as you can imagine its calculation is data-intensive (you need to know the holdings of thousands of mutual funds in order to make any conclusions).
To continue reading this article please login (at the right) or click here to learn more about accessing our archives.
Related Posts
- Study finds that before “swinging for the fences” HF managers are influenced by several factors
- New study of mutual fund alpha shows that what-goes-around-comes-around
- Study finds many hedge funds simply hold back liquidity to power returns
- Study finds secondary HF markets can predict future fund returns
- New study finds 130/30 outperforms long-only in back tests
“This is a pretty significant finding since it suggests that markets are not perfectly efficient – maybe not even “semi-efficient”. In fact, they appear from this study to be predictable based solely upon prior returns.”
– Not to be argumentative but the question of markets not being efficient isn’t truly in doubt, is it?
“The predictive relationship between r-squared and alpha also works in reverse, say the authors of this paper. Funds with high alpha performance one year tend to have a low r-squared values the next. The authors suggest this may be because…“…funds with exceptionally good performance in one year tend to take idiosyncratic bets in the following year because they can ‘afford it’ given past success and given that funds are usually judged over a number of years…” ”
— this is essentially the same as the neuroeconomically demonstrated ‘house money effect’. The actual psychological mechanism in place is a greater feeling of confidence that changes the manager’s perception regarding the risk of the next bet. Yes they can afford it but it is the feeling of confidence in their judgment that allows the additional risk to be taken.