Risk bookkeeping with Monte Carlo aggregation is still just risk bookkeeping Risk management systems in #EnterpriseRiskManagement differ fundamentally from risk management systems at banks and insurers. In ERM, risks are assessed by experts outside the system, then entered into a system and simply stored there until they are added up through a simulation. The system is dumb, and the expert has miraculous powers. He can supposedly grasp and evaluate, in his head, all the many uncertainties that influence a risk. If he can really do that, he could just as well go ahead and do the company-wide aggregation himself. That would not be any more complicated. Most of the time, however, the result is just a distribution built on the good old best case, most likely case, and worst case assessment, the way it was done in the era before computers arrived. There is no validation at all. This approach sits squarely in the garbage-in, garbage-out trap. A Monte Carlo simulation for the further aggregation changes nothing about that. A banker or actuary would not expect to have to tell the risk system how big the risks are. On the contrary, he would assume that the ability to assess risk is precisely the purpose of the system, and that he provides factual information as input, such as the type and size of transactions and customers. The #EnterpriseRiskEvaluator is therefore open, alongside expert assessments, to new data streams and facts, and can actively assess risks according to your specifications. Abstract risks become concrete risk drivers here, ones that are connected to each other and for which you can find data that can be measured and validated. The individual risk is thus already the result of an 'aggregation before the aggregation' that you design yourself. I look forward to your opinions and discussions on this important topic. https://lnkd.in/e4Sk7giE
This post was originally published on LinkedIn. View and join the discussion there