Hierarchical Risk Budgeting

Drawing on a recent contribution to the literature on risk budgeting (RB), the author investigates an RB approach based on the filtered similarity matrix generated by hierarchical clustering. Like the hierarchical risk parity (HRP) approaches, the resulting RB approach, which the author calls hierarchical RB (HRB), has attractive properties such as visualization, flexibility, and robustness. Relying on its flexibility and robustness and exploiting its two-step formulation, the author then describes how HRB can be extended to incorporate the expected returns on assets without additional constraints to control estimation errors. Finally, the author tests the empirical performance of HRB against that of the classical HRP. The findings show that HRB is a promising alternative to the classical HRP, particularly when it takes into account the expected returns on assets. The author checks the robustness of these findings using block bootstrapping.