Monte Carlo Simulation Engine
TERM also includes a high speed Monte Carlo simulation engine which allows users to calculate the full loss distribution curve of the earnings at risk for different risk scenario combinations and settings.
The application calculates the quantity by which net income might change due to the risk exposure of the organization over a year. A Monte Carlo simulation can be used to model the distribution of the earnings variability over a time period by running multiple simulations. Stochastic variables are used for the occurrence, size of loss/gain and outcome (loss, no loss, gain) of the risk scenarios defined by the experts.
The correlations between pairs of individual risk scenarios can be included in the quantification in order to model their interdependencies. Many risk scenarios are independent, i.e. correlation coefficient equals zero. However, some risk exposures are dependent and correlated to different degrees. Some exposures leading to an event can trigger other events (positive correlation), leading to a super-additivity condition of the total risk exposure. Vice versa some exposures leading to an event can exclude other events or decrease its risk level (negative correlation) which would lead to a sub-additivity condition of the total risk exposure and be a natural hedge for a company. Additionally, some risk exposure dependencies are asymmetrical; earthquakes might lead to fire (when gas lines are ruptured, releasing gas, or power lines brought down, causing arcing and sparks), but a fire would not lead to an earthquake. Asymmetrical correlations and dependencies can be dealt with by using copulas distribution functions and the modeling of the total risk exposure can be straight forward by using Monte Carlo Simulations.