Predictive Blends : Fundamental Indexing Meets Markowitz

When constructing a portfolio of stocks, do you turn a blind eye to the firms' future outlooks based on careful consideration of companies' fundamentals, or do you ignore the stocks' correlation structures which ensure the best diversification? The Fundamental Indexing (FI) and Markowitz mean-variance optimization (MVO) approaches are complementary but, until now, have been considered separately in the portfolio choice literature. Using data on S&P 500 constituents, we evaluate a novel portfolio construction technique that utilizes the benefits of both approaches. Relying on the idea of forecast averaging, we propose to blend the two previously mentioned techniques to provide investors with a clear binocular vision. The out-of-sample results of the blended portfolios attest to their superior performance when compared to common market benchmarks, and to portfolios constructed solely based on the FI or MVO methods. In pursuit of the optimal blend between the two distinct portfolio construction techniques, MVO and FI, we find that the ratio of market capitalization to GDP, being a leading indicator for an overpriced market, demonstrates remarkably advantageous properties. Our superior results cannot be explained by classic asset pricing models