CfS Afternoon on Risk, Noise, and Extremes, Edinburgh

https://centreforstatistics.maths.ed.ac.uk/cfs/events/cfs-afternoon-risk-noise-extremes

**Abstract**: The World Bank has identified parametric (or 'index') insurance as an underused tool for managing risk and reducing poverty. In outline, such an insurance contract might read 'pays $100M the first time a large explosive eruption occurs at an Indonesian volcano in the next year'. The challenge is to calculate the fair price of such a contract. This is one example of the type of problem I consider: given a catalogue of events from cases that are similar but not identical, compute the probability distribution of the number of events of a specified subset, over a specified future time-period. Technically, this appears to be a difficult calculation, given that it involves a convolution over non-identical and non-independent random variables. However, a carefully chosen set of modelling choices makes this a computationally tractable problem that scales well to thousands or millions of cases. This means it can be used for volcanic risk screening (hundreds of cases), and also in many instances of everyday insurance, such as car accidents (millions of cases).