Stochastic frontier analysis for investment-fund risk–return frontiers

Data envelopment analysis (DEA) is a method for evaluating economic efficiency that has been adapted to compare investments with different risks and returns (Gregoriou and Zhu, 2005). Traditional DEA does not allow for risk reduction through diversification. Recently Lamb and Tee (2012a) developed DEA method that does, and then (2012b) showed how to obtain statistical comparisons among investment efficiencies. However, their method uses a piecewise linear approximation to the efficient frontier and gives biased efficiencies. An alternative approach is stochastic frontier analysis (SFA). SFA has the benefit of modelling a smooth approximation to the frontier but in its usual form is not correctly specified fro the investment-fund risk–return frontier. This project will use structural models to develop a new SFA methodology and investigate its statistical properties and whether it can be combined with DEA.
A reasonable knowledge of statistics or econometrics is required. Knowledge of statistical modelling software such as R, Stata or possibly Mathematica is desirable.