“People talk about black swans but they don’t talk about robustness, which is the real lesson of the black swans.”
—Nassim Taleb, author of The Black Swan, in a April 13, 2011 interview as part of Wharton Business School’s Goldstone Forum.

The terms ‘“parameter risk”‘ and “‘model risk”‘ fill at least one important need in the actuarial lexicon: their frequent use reminds us that the models on which we base financial decisions could be materially wrong. While the practice of ERM focuses on quantifying the various risks threatening an insurance enterprise, model risk yet remains and poses something of a paradox. Can a model assess itself? While it may quantify volatility in a risk process, what is the proper approach to quantifying uncertainty in the model? More importantly, how can we make better decisions in the face of model risk?