Abstrakt
The new class of data envelopment analysis (DEA) models with diversication suitable for nancial applications has gained a special attention in recent years. These models take into account possible dependencies between the considered asset returns which were not considered by the standard DEA models.
Several solution methods have been proposed for the resulting nonlinear problems, however, these methods are restricted to discrete distributions only. In this paper, we consider a multivariate continuous distribution of the random returns and use the sample approximation technique to solve the DEA problems approximately.
We focus on the quality of the approximation. Moreover, we introduce a one-step procedure where inputs for a benchmark need not to be computed before solving the DEA model.
The new methods are applied to access eciency of selected assets from US stock market.