A new stochastic dominance approach to enhanced index tracking problems

Enhanced Index Tracking is the problem of selecting a portfolio that should generate excess return with respect to a benchmark index. Here we propose a large-size linear optimization model for Enhanced Index Tracking that selects an optimal portfolio according to a new stochastic dominance criterion and we devise an efficient constraint generation technique to solve such a model. We then compare, on several well-known and publicly available financial data sets, the performances of the portfolios selected by our model to those of the portfolios obtained with other stochastic dominance approaches. The results seem to confirm the practical usefulness of stochastic dominance for portfolio selection.