Operational Risk Model

Prescio has developed an in-house operational risk model that merges a Bayesian methodology with the Loss Distribution Approach (LDA).

This modeling approach allows for the seamless combination of internal data, external data, and scenario analysis in the model, as required within the Basel II framework. Prescio uses this in-house model as a comparison when externally validating a bank’s operational risk framework.


Operational Risk Model Outline:

  • The Loss Distribution Approach (LDA) is followed with Bayesian inference used to derive the frequency and severity distributions.
  • The parameters in the frequency and severity distributions are modeled by “posterior distributions” which are calculated via Bayesian inference.
  • The posterior distribution is the product of the prior distribution and the likelihood functions.
  • The prior and posterior distributions have the same form, which by definition is the conjugate prior of the distribution being modeled (whether frequency or severity).
  • Structured scenario data is used to derive the prior distribution.
  • Internal and external data are used to calculate the likelihood functions in the posterior distribution.


Related Links