Paul Larsen is Head of Data Science with Allianz Global, Corporate and Specialty insurance in Munich, Germany. He was previously responsible for the Allianz Group's operational risk Solvency II model, and before that worked for Deutsche Bank in Berlin on portfolio modeling, with a focus on operational risk.
His PhD is in mathematics (algebraic geometry) from Humboldt University in Berlin, and he also has two degrees from Oxford University as a Rhodes Scholar.
The aim of this paper is to systematically investigate the stability of operational value-at-risk (OpVaR) models when fitting heavy-tailed distributions to the relatively small sample sizes found in operational loss data.
In this paper, the author studies how asymptotic normality does, or does not, hold for common severity distributions in operational risk models.