Applied mean-ETL optimization in using earnings forecasts

In this article, we apply the mean-expected tail loss (ETL) portfolio optimization to the consensus temporary earnings forecasting (CTEF) data from global equities. The time series model with multivariate normal tempered stable (MNTS) innovations is used to generate the out-of-sample scenarios for the portfolio optimization. We find that (1) the CTEF variable continues to be of value in portfolio construction, (2) the mean-ETL portfolio optimization produces statistically significant active returns, and (3) the active returns generated in the mean-ETL portfolio with CTEF indicate a statistically significant stock selection.