A Multiobjective Multiperiod Efficient Portfolio Selection Approach for Positive and Negative Returns Using RDM-DEA Under Credibilistic Framework

Abstract

There are abundant applications of data envelopment analysis (DEA) for efficient portfolio selection, though in a single-period environment. However, portfolio selection is a multiperiod process involving long investment horizons. Subsequently, owing to changing market conditions, investors often rebalance their portfolios based on the performance of each asset and the portfolio as a whole. Besides, assets' performance constituting the multiperiod portfolio changes from one period to the next. Consequently, assets may yield positive or negative returns in one or more periods. Hence, it is imperative to deal with negative returns in a multiperiod framework for a multiobjective multiperiod efficient portfolio selection (MOMPEPS). To this end, this article proposes a two-stage credibilistic MOMPEPS approach. A multiperiod range directional measure DEA model is employed in Stage-I for evaluating the period-wise and overall efficiency of assets having both positive and negative returns. In Stage-II, a MOMPEPS model is proposed for maximizing the terminal wealth, overall efficiency, and minimizing the cumulative risk of the portfolio subject to several realistic constraints. The max-min approach solves the proposed model and presents several efficient portfolios for different investment schemes characterizing investors' attitudes. A comprehensive numerical illustration demonstrates and validates the proposed approach.