Framework for Developing Attribution Models. Symmetrical Arithmetic and Geometric Attribution

The book introduces an efficient practical framework for development of attribution models, which present one of the major analytical vehicles in investment analytics. The framework is based on robust mathematical and business concepts. First, the author gives a brief overview of existing attribution models from the mathematical and business perspectives. He shows that the present attribution models have certain principal inconsistencies, and provides explanation why they are inherent to existing models. Then, based on this analysis, several concepts and criteria are introduced in order to address the discovered problems at the core level. For instance, one criterion states that attribution models should satisfy a symmetry principle, that is, the results of attribution comparison of a portfolio to a benchmark should be the same as the results of comparison of a benchmark to a portfolio, which is not the case with the present attribution models. On the basis of the proposed framework, two new arithmetic and one geometric attribution model that are free from the drawbacks of existing approaches were developed and thoroughly validated. The author shows that these models deliver objective attribution parameters meaningful for investment business applications; both at the total level, and at the levels of separate financial instruments and their groups.