Abstract
The paper deals with out-of-sample efficiency of particular portfo- lios with respect to the second order stochastic dominance (SSD). Firstly, using Conditional Value at Risk (CVaR) as the measure of risk, we compute three mean-risk efficient portfolios from the mean-CVaR in-sample efficiency frontier.
These portfolios are identified as the optimal solutions of the CVaR minimizing problem under the condition on minimal required portfolio mean return. Three levels of minimal required portfolio mean return are considered.
Then we test out-of-sample efficiency of all these portfolios with respect to the second order stochastic dominance. We apply this procedure to US stock market daily data 1927-2014 such that we use one year moving window period for both in-sample and out-of-sample analysis.
Hence we identify 87 in-sample sets of three mean- CVaR efficient portfolios and we test SSD efficiency of all these 261 portfolios. Moreover, we analyze the out-of-sample efficiency (measure of SSD inefficiency) for each minimal required level of portfolio mean return separately.