ABSTRACT
This paper empirically compares the sampling properties and out-of-sample performance of several reputed shortfall-based portfolio selection rules by applying the bootstrap and k-fold cross validation to two distinct data sets spanning five decades. The results suggest that a simple extension of the venerable Safety First rule offers the best overall performance, especially when methodological simplicity is desired.
Acknowledgments
The author is grateful to David Fuller and M. Kevin McGee for helpful comments and useful suggestions.
Disclosure statement
No potential conflict of interest was reported by the author.
Notes
1 See Hodder, Jackwerth, and Kolokolova (Citation2015), Ameur and Prigent (Citation2018) or Chen and Lu (Citation2019) for other applications of reward-risk ratios.
2 See Haley, (Citation2017) for a discussion of k-folding in a finance setting.