Comparing and selecting performance measures for ranking assets

Within an asset allocation framework, when the number of assets is larger than the sample dimension, mean-variance approaches cannot be used due to the limited number of degrees of freedom. In such a situation, performance measures could be used to rank assets, and then select a subset of them for further analysis. However, the financial economics literature proposes dozens of measures, and there is thus a problem: which measures should be considered? Some authors already discussed this topic. We extend the current literature by enlarging the set of analyzed measures and also by exploiting the possible dynamic evolution of rank correlations. Our analysis is mainly empirical, based on the S&P 1500 constituents, and includes an example of the optimal combination of performance measures for allocating an equity portfolio.