Effectiveness of earnings forecasts in efficient global portfolio construction

We analyze the effectiveness of using fundamental variables of earnings forecasts for constructing mean-variance efficient portfolios. The performances of the Markowitz mean-variance optimal portfolios are examined by selecting stocks based on the consensus temporary earnings forecasts (CTEF) data. An empirical analysis on both US domestic equities and international equities is conducted for the period 1997-2010, and we find that the CTEF variable is a statistically significant factor in generating portfolios with active returns over benchmark portfolios.