Why Analyze Hedge Funds with PMVD?
About PMVD
PMVD is based on bargaining theory. Imagine that factor owners bargain to get as much explained variance credited to their factors as possible. Assume that owners can form coalitions and that the bargaining power of any coalition is the joint marginal contribution to explained variance of its factors. Then one particular bargaining solution, the proportional value, divides up explained variance in a very desirable way.
PMVD is consistent with statistical theory. A factor with zero contribution to explained variance receives zero variance share. It is the only variance decomposition method with this property.
PMVD variance components are a measure of relative importance. Relative importance is not completely determined by factor betas or their statistical significance. As you will see, it is possible that one variable is statistically more important than another even though it has a lower statistical significance level. Relative importance, as measured by PMVD, takes a factor’s joint contributions to explained variance into account.
PMVD does not require statistical constraints to obtain understandable results.
PMVD results can be evaluated more quickly and are easily evaluated graphically.
PMVD components are non-negative. They are reported signed by the sign of the factor beta in order to facilitate interpretation.
Copyright © 2006 Prism Analytics