Analysts’ Dividend Forecasts, Portfolio Selection, and Market Risk Premia

Abstract

Today, there is no doubt that modern portfolio selection theory was initiated by the famous contribution of Markowitz [5]. However, even more than fifty years later the practical relevance of Markowitz’s work lies far behind the theoretical impact of his ideas. Practical applications of the Markowitz approach are mostly impeded by the necessity of the adequate estimation of expectation values, variances, and covariances of security returns. While historical estimations for variances and covariances work quite satisfactorily, return realizations are only a poor proxy for actual expected future returns. We want to examine the efficiency of portfolio management decisions based on estimated expected returns derived from analysts’ dividend forecasts (thereby allowing for non-flat term structures of interest rates and German tax rules) in comparison to nine other portfolio selection strategies. Moreover, we analyze consequences of expectation biases in dividend forecasts and estimate market risk premia.