ERM Glossary: Extreme event
[this page | pdf | references | back links | custom searches]
Extreme events are outcomes that lie in the tail of the
total loss or payoff distribution of the enterprise. As explained in Kemp (2010), what is or
is not an ‘extreme event’ always implicitly requires some prior probability
distribution against which observed outcomes are being compared.
Measuring and modelling such risks is difficult, not least
because there is usually relatively little data available relating to the tails
of the distribution. For some types of risk, modelling of extreme events may
benefit from the use of Extreme
Value Theory (EVT). For further information on EVT see here.
Contents | Prev | Next