Active risk sensitivity to views using the Black–Litterman model

Abstract

In this article, we introduce a new measure for the marginal contribution of each view to the expected tracking error volatility (TEV). Our findings are obtained in the Bayesian framework proposed by Black and Litterman (1990, 1992) considering both a world with and without constraints in the optimization process. We believe that the issue of TEV sensitivity to the views can be of considerable interest to the investment management community for several reasons as it (1) provides the portfolio manager with a method for revising the portfolio consistently with a given TEV constraint or for better understanding its impact on asset exposures; (2) makes investment specialists responsible for the generation process of the views; (3) creates the basis for a procedure to examine a request from the asset managers to expand the relative risk budget. We provide an empirical investigation illustrating how our new measure can be implemented to modify a set of given views to achieve a TEV goal.