Risk management and portfolio selection using α-stable regime switching models

This article tries to enhance traditional distribution paradigms for modelling asset returns by considering an α-stable regime-switching model. Our approach is to perform an empirical test of the α-stable regimeswitching model against other common methods in two settings: in risk management and in portfolio selection. Our empirical study will show that the model is better suited than Gaussian and Gaussian regimeswitching models to measure risk accurately. A portfolio optimization case study for a traditional stocks and bonds investor is pursued. In this study, the model leads to less risky and more diversified portfolios. In particular, the model avoids outsized losses in times of crisis and thus leads to a better (adjusted) Sharpe ratio and Omega.