Abstract
The paper deals with mean-risk models using Conditional Value at Risk (CVaR) as a measure of risk. Firstly, using daily data of returns of 49 US industrial representative portfolios within one year, the optimal in-sample portfolios for years 1970 - 2011 are computed.
Then the out-ofsample performance of these portfolios is analyzed using annual returns from years 1971 - 2012. The analysis is done for various values of risk parameter and the out-of-sample optimal value of risk parameter is identified.
The out-of-sample optimality is considered with respect to various criteria: maximization of mean out-of-sample return, minimization of variance of outof- sample returns and minimization of CVaR of out-of-sample returns. Finally, the out-ofsample returns are compared using second-order stochastic dominance relations and efficiency.