Experience and assumptions
How much should you believe your own data?
Credibility is taught as a formula and applied as a reflex. The harder question is what sits on the other side of the blend, and most assumption bases never say.
Insights
Writing on assumptions, reporting, modelling, analytics and review — the working problems rather than the settled theory.
Credibility is taught as a formula and applied as a reflex. The harder question is what sits on the other side of the blend, and most assumption bases never say.
Most actuarial reporting fails at the last mile. The numbers are right, the refresh is automated, and the person it was built for still asks for it in an email.
The same walk serves the Prudential Authority and the executive committee. It rarely serves both well without being rebuilt for the second audience.
An order of operations for model remediation: what to test first, what to document, and the point at which rebuilding becomes cheaper than repairing.
Independent review has a predictable shape. Knowing what an experienced reviewer opens first is the cheapest way to improve the report before it is sent.