Robust Portfolio Optimization with Multiple Experts

The success of quantitative approaches to portfolio choice crucially depends on the considered return model. Experts however do not agree on which return model is most appropriate. This controversy about the model specification introduces uncertainty in the optimal portfolio choice. We will not meddle in the discussion on which model specification is most appropriate. Instead we consider the advice of various experts and aim for a portfolio choice which is robust to the advice of all the experts. More specifically we consider the portfolio choice with maximal performance for the most pessimistic advice. We examine the effects of robust mean-variance portfolio choice: the difference with non-robust portfolio choice and its mean-variance performance. We test the robust approach empirically in the context of international portfolio choice with an investment set consisting of 81 portfolios from 9 European countries. Advice regarding the appropriate return model will be obtained from alternative experts advocating the CAPM, the international CAPM, the international Fama and French factor model. A bootstrap experiment demonstrates the merits both in terms of expected performance as well as worst case performance of the robust approach.