Stochastic loss reserving

Whereas deterministic models produce single-point estimates of future liability, stochastic loss reserving models incorporate the random variation of the various assumption parameters over time, and thereby produce a distribution of the outstanding claims liability. Stochastic models can take the form of any of the typical deterministic approaches (i.e. a stochastic PPCI, stochastic PPCF, etc.) and are generally constructed as generalised linear models (GLMs). The family of stochastic models can be illustrated as follows: