PySDTest: a Python Package for Stochastic Dominance Tests
We introduce PySDTest, a Python package for statistical tests of stochastic dominance. PySDTest can implement the testing procedures of Barrett and Donald (2003), Linton et al. (2005), Linton et al. (2010), Donald and Hsu (2016), and their extensions. PySDTest provides several options to compute the critical values including bootstrap, subsampling, and numerical delta methods. In addition, PySDTest allows various notions of the stochastic dominance hypothesis, including stochastic maximality among multiple prospects and prospect dominance. We briefly give an overview of the concepts of stochastic dominance and testing methods. We then provide a practical guidance for using PySDTest. For an empirical illustration, we apply PySDTest to the portfolio choice problem between the daily returns of Bitcoin and S P 500 index. We find that the S P 500 index returns second-order stochastically dominate the Bitcoin returns.
READ FULL TEXT