Why Analyze Hedge Funds with PMVD?
Existing Returns-Based Methods
Let’s review current returns-based techniques.
Factor models produce factor betas and statistical significance levels. The significance of factor betas is often low because of high levels of correlation between factors. Precision may be improved by requiring that betas sum to one. This is the no-arbitrage constraint. It assumes that futures positions are fully collateralized and that options are hedged asset positions.
Style analysis imposes the additional constraint that factor betas are non-negative. This assumption is untenable when profiling against conventional markets. It is problematic even when using hedge fund benchmarks. It becomes impossible to have a beta greater than one against any factor.
Covariance decomposition is based on betas and factor covariances. These components can be unstable due to factor colinearity. A factor’s covariance component can be zero even though it is statistically significant. Components can be negative and signs may not agree with the signs of factor betas.
Copyright © 2006 Prism Analytics