In Parts I to IX we have presented the tools required to build a coherent asset allocation in the presence of stress events. In the last part of the book we look in detail at the allocations obtained using this procedure from three different perspectives.

In Chapter 26 we obtain the optimal allocations for a simple but realistic problem; we discuss the properties of the solution; and, what is most important, we try to understand the intuition behind the results.

In Chapter 27 we explore how well the many approximate numerical techniques that we have described in the previous chapters actually work. As we shall see, the answer is that they are surprisingly effective, and that they can therefore greatly reduce the computational burden of the optimization.

In Chapter 28 – one of the most important of the book – we look at the sensitivity of the results to the unavoidable uncertainty in the inputs. We shall discover that the allocations are very sensitive to the precise values of the expected returns. We hasten to stress that the high sensitivity of the results to the expected returns is not brought about by the Bayesian-net methodology – if anything, the latter goes some way towards stabilizing the results. But, as we shall see, there is something rather deep a bout asset allocation in general, and this lack of stability in particular, that these results will bring to light.