Abstract

Application of Data Envelopment Analysis (DEA) in selection of portfolio is for occasions when we are to consume rate of a capital in a financial market can be useful. In enforcement of this application, the most significant factors influencing on the performance evaluation of the investment companies are mean of the profit achieved from investment and it's variance. One of the discussed assumption in performance evaluation of portfolio can be in such a manner that the more the mean of the resulted profit and the less the variance, the better will be the performance. For this reasons mean as output and variance as input have been taken into consideration in the technique of DEA. But, since efficiency in DEA results from division of output by input, thus, variance may overshadow the mean, and a DMU which has a very low mean and little variance is to be placed better than a DMU which has a high mean and relativelyhigh variance. For this reason, in this article, a criterion is presented when variance as an input is supposed, but it is placed under shadow of mean. In other words, the first priority is evaluation via mean, but this evaluation must not be absolute priority. Supposing that there are n portfolios with equal inputs (purchase price) and various outputs (mean and variance) and considering variance as an undesirable output, we create a change in the mean with aid of the suggested relationship, and, then, technique of DEA will be enforced. Finally, this technique is to be employed on a real data set.
