The treatment presented in the previous chapter makes extensive use of the notion of ‘events’. Where do ‘events’ come from? How are we choose them? How can we relate them to the positions in our portfolios?

In this chapter we answer these questions. In a nutshell, we find that, when it comes to building a causal Bayesian net for portfolio management purposes, a useful way to organize our thinking is to make use of four clearly distinct components:

1. root event(s);

2. transmission channels;

3. changes in the market risk factors that affect the value of a given portfolio;

4. deterministic mappings to portfolio profits and losses.

The task faced by the expert is how to connect the root events with the changes in the market risk factors that affect a given portfolio. The connection is mediated by causal transmission channels, i.e., specific events triggered by the root event(s).

Of course, there could in principle be a variety (and perhaps an infinity) of transmission channels whereby a given root event may affect a set of market risk factors. The skill of the expert lies in her ability to identify the most plausible linkages given the specific context attaching to the chosen root event. ‘Most plausible’ is clearly an imprecise term, and good results tend to be the outcome of vivid, yet restrained, imaginations. For reasons that we explain later in the chapter, we like to call these connecting transmission channels ‘paths of low resistance’.