Portfolio optimization using data envelopment analysis: Exchange-traded funds portfolio selection and allocation

This study aims to present a model that incorporates risk and return benchmarks, with the intention of optimizing the performance of investment portfolios. Consequently, this study employs two Data Envelopment Analysis (DEA) models, specifically the slacks-based measure (SBM) of efficiency and super-efficiency SBM (super-SBM), to identify potential Exchange-Traded Funds (ETFs) for investment. A total of 34 ETFs were selected for the application of these models. An inherent challenge in utilizing DEA models is the management of negative, zero, or near-zero inputs or outputs, which have the potential to influence the model results or render them unacceptable. Within the financial context, these challenges manifest in cases such as negative Sharpe ratios or near-zero returns, so the utilization of the CCR model is not applicable in the case of negative variables, while the BCC model is capable of accommodating negative variables under specific circumstances. Furthermore, the presence of near-zero or zero variables renders the results of the AP model unstable or infeasible. The results of this study indicate that the SBM not only performs satisfactorily under the aforementioned challenges but also effectively identifies potential ETFs, and indicate that seven of the 34 ETFs were identified as efficient and suitable for investment. Moreover, the utilization of super-SBM assists portfolio managers in optimizing fund allocations. The proposed framework provides a systematic approach to ETF selection and portfolio allocation in real-world investment settings.