A Comprehensive mean-maverick-variance portfolio selection model based on the OVG aggregated game cross-efficiency approach

This work makes two important contributions: first, we present the aggregated DEA game cross-efficiency concept and use the ordered visibility graph (OVG) approach to determine weights. Further, we integrate this OVG game cross-efficiency approach with a multiobjective mean-maverick-variance (MMV) portfolio selection model. In the proposed work, the OVG game cross-efficiency value is treated as a return characteristic, and the Maverick index & variance of cross-efficiencies as a risk characteristic. We improve game cross-efficiency as a tool for portfolio selection by introducing the OVG game cross-efficiency scores, which provide a set of unique Nash equilibrium game cross-efficiency scores and faster convergence. The empirical component uses a database of 50 companies listed on the National Stock Exchange, India, to show that our approach can be helpful in diversified portfolio selection. When the cardinality constraint is set to 10, the results show substantial variation across key metrics with returns ranging from 0.082 to 0.1475, efficiency scores from 0.8152 to 0.3652, maverick values from 0.0453 to 0.1231, and variance from 0.0035 to 0.0248. This wide dispersion across indicators reflects a higher level of diversity compared to that observed in existing models.