The risk that arises due to an error in financial risk measurement and valuation models. Errors give rise to cases of inadequacies and shortcomings in measurement of risk and the valuation models applied, which in turn result in financial losses to an entity.

A wide variety of different types of fundamental models used across the industry includes the Black-Scholes option pricing model that is premised on the dynamic processes and interrelationships between different variables (determining the price of an option). Another widely used set of models are “statistical models” that are designed to identify statistical relationships between variables, mainly focusing on the correlation between variables. Models are applied to help an entity take financial decisions about its loss limits and risk budgets. Notwithstanding, models are simplified set-ups, aiming to depict and predict more complex behavior of variables. Certain proportion of error and discrepancy is considered a normal outcome.

Model risk stems from a variety of reasons or sources, primarily 1) incorrect model specification (underestimation of relevant risk factors, wrong specification stochastic processes underlying the model 2) incorrect model application (using the wrong or old-fashioned model for the problem at hand) and 3) implementation risk factors exemplified in a) incorrect calibration of model parameters, b) programming errors or c) problems with data sufficiency and lack of up-to-date model input information.