Why Analyze Hedge Funds with PMVD?
PMVD Limitations
PMVD is a statistical procedure. You should approach any statistical procedure expecting some noise. PMVD helps reduce the effects of statistical noise but cannot eliminate it.
Further, all returns-based methods are inherently backward rather than forward looking. Even holdings can only tell you about the present. However, returns-based methods have a good dose of the past mixed in as well.
Also, variance decomposition is not tail risk decomposition. However, a skilled analyst can often make useful tail risk inferences.
We think about volatility in standard deviations. Variance is its square. Working with variance takes some getting used to.
Finally, PMVD components are not additive. Thus, the portfolio variance component of a factor many not be equal to the weighted sum of the components for individual funds.
Copyright © 2006 Prism Analytics