The idea is pretty simple, which I’ll outline below. (I can do so, as the materials are already in the public domain.)
If you want the details, check out the formal paper at SSRN. For a more accessible introduction, I’d recommend the video below that was taken (and won an award) at the International Congress of Actuaries in 2014.
The gist of the idea
In short, a key mantra of investment advisors is that asset owners should ignore the performance of their asset manager. Whether the manager has outperformed for you or not, it is claimed, is irrelevant to how you should expect them to perform in the future. However, a key idea in statistics, called Bayesian thinking, suggests the opposite: that you should use any new information you get about something to update your thinking on it.
In this work, I therefore used Bayesian thinking to understand the advice that asset owners should get. It turns out that this advice makes common sense, particularly when paired with an understanding of whether the asset manager has some longer-term cyclicality in their performance.
Here’s that video that I mentioned: