Why Independent Data Validations are Vital for Model Risk Management

In the world of model risk management (MRM), the integrity of the data within a model is extremely important. Given that models rely heavily on data to generate predictions or inform decisions, any inaccuracies or inconsistencies in the data can significantly undermine the reliability and effectiveness of the models. This is where independent data validations come into play.

The Importance of Independent Data Validations in Model Risk Management (MRM)

Independent data validations play a pivotal role in MRM by providing an objective assessment of the data inputs used in models. This process involves a thorough examination of this data to ensure accuracy, completeness, and appropriateness. A review of both data inputs and outputs provides critical safeguards against potential errors and biases that could compromise the integrity of the models.

An independent data validation process involves an objective assessment of the data inputs and outputs used in models, free from influence or bias from the model development or implementation teams. Regulations such as the Supervisory Guidance on Model Risk Management from the Office of the Comptroller of the Currency (OCC) and the Federal Reserve, later adopted by the Federal Deposit Insurance Corporation (FDIC), call for an independent validation of identified models. This means there should be appropriate segregation of duties, particularly between those individual responsible for the development of the model from those charged with validating the model.

Effective MRM Practices

When the same individual or team is responsible for both data validation and optimization reviews, there is a risk of inherent bias or oversight, as they may be inclined to prioritize certain outcomes or overlook potential weaknesses in the model. An optimization review (often referred to as a tuning or calibration review) offers management decisions as to how certain parameters should be set within a model. By segregating these duties, organizations can maintain the integrity of the validation process and uphold the principles of sound model development practices.

Management should ensure that different individuals or teams are responsible for data validation and optimization reviews. Meaning, those responsible for determining the parameters within the system should be separate from those performing the validation. This segregation of duties ultimately contributes to the overall effectiveness and credibility of MRM practices, strengthening the resilience of the organization in an increasingly complex and dynamic environment. If organizations are relying on a third-party for model validation or optimization services, management should ensure the same vendor is not performing both the optimization and validation to ensure proper independence.

Interested in speaking with our trusted model risk management professionals? Reach out to a member of our team today!