What the Black–Litterman approach tries to achieve

The methodology we present in this book has obvious links with the Black–Litterman framework: in a nutshell, Black and Litterman suggest how to construct an optimal portfolio once the investor's subjective views about expected returns are supplied. The portfolio we construct is a function of the subjective views about asset returns and codependencies. Black and Litterman elicit subjective information about where the distributions of returns are centred. We also require subjective information about what will happen in the tails. Since in our approach the input subjective views are more complex, it is essential to work through their logical consequences in a self-consistent and structured manner. This we do using Bayesian nets.

In order to move beyond the soundbites, and to appreciate more clearly what the links (and the differences) between their approach and our approach are, we provide in this chapter a concise presentation of the Black–Litterman model. As in the case of the Markowitz model, our treatment is not meant to be a balanced introduction to the Black-Litterman approach, but a rather idiosyncratic presentation focused on its subjective aspects, and on the stability of the asset allocations that it produces.

We saw in the previous chapter that one of the problems associated with the Markowitz approach is the instability of the solutions (allocations) it can produce.