A Rule of Thumb for the Optimal Number of Runs in Monte Carlo Simulations

Abstract

In general, the understanding of risk could be distinguished in three possible ways; known-knowns, known-unknowns and unknown-unknowns. The known-known risk would certainly refers to the one we know we know, the one we accept. This falls under the daily operations in an risk organization. Further, there are several known-unknowns in risk management. Thoose can be accounted for in some extent in our analysis and strategic decisions, by e.g diversification, investing and improved project management. The unknown-unknown risk would refer to the information of risk we are not aware of, and even worse, are not aware that we are not aware of it. It is almost impossible to mitigate the unknown-unknown risks, however by understanding the known-unknown risks would certainly reduce this uncertainty. This paper discusses one of the most famous known-unknown risk in operational risk modelling, the optimal number of Monte Carlo simulations. We develop and reveals a formula that will upfront make the known-unknown process to a known-known fact.

Keywords: Monte Carlo simulation, Coefficient of Variation, Known-unknown risk, Operational risk

JEL Classification: C15, C63