Mean-ETL Optimization in HorseRace Competition

Abstract

In this paper, we present the methodology and results of the portfolios submitted to the HorseRace competition. The nine portfolios were constructed by applying the Mean-ETL optimization approach. The Mean-ETL optimization approach uses three fundamental variables (CTEF, MQ, and REG10) and three stock universes (GL, XUS, and EM), with each of the three fundamental variables applied one at a time to one of the three universes. This study assesses the return of the nine portfolios, and we report that all of these Mean-ETL portfolios produce positive active returns and most of them are statistically significant. Additionally, MQ variable is found to be the best among these three variables in the Mean-ETL portfolio construction.