Statistics Seminar: Allen -- A Multi-Criteria Portfolio Analysis of Hedge Fund Strategies

This paper features a tri-criteria analysis of Eurekahedge fund data strategy index
data.  We use nine Eurekahedge equally weighted main strategy indices for the portfolio
analysis.  The tri-criteria analysis features three objectives: return, risk and
dispersion of risk objectives in a Multi-Criteria Optimisation (MCO) portfolio
analysis.  We vary the MCO return and risk targets and contrast the results with four
more standard portfolio optimisation criteria, namely the tangency portfolio (MSR), the
most diversied portfolio (MDP), the global minimum variance portfolio (GMW),
andportfolios based on minimising expected shortfall (ERC).  Backtests of the chosen
portfolios for this hedge fund data set indicate that the use of MCO is accompanied by
uncertainty about the a priori choice of optimal parameter settings for the decision
criteria.  The empirical results do not appear to outperform more standard bi-criteria
portfolio analyses in the backtests undertaken on our hedge fund index data.  

Keywords: MCO, Portfolio Analysis, Hedge Fund Strategies, Multi-Criteria Optimisation,
Genetic Algorithms.  

JEL Codes: G15 , G17 , G32 , C58 , D53.