The regulation represents a key point to consider given that regulators are always more interested in model risk management after the last financial crisis. The previous and current regulation on model risk is increasing the level of the requirements for a MRM framework to make it sure it covers the end-to-end process in a robust and coherent way. A key point for the regulators is the quality and completeness of the model documentations. For this reason, a financial institution should establish processes and policies to identify and define roles and responsibilities in the documentation management when approving or making changes in the respective model documents. Last but least, it is essential to document all the technical decisions made (e.g. calculation method, adequacy for a specific product, overlay etc).
Review of the validation process due to the huge evolution of machine learning (ML) and big data. The possible outcome is that the manual process will be replaced by automated test suites to perform the more time-consuming tasks on data input review, calibration on a vast amount of data at the same time. In addition, financial institutions should consider how to revise their model risk policy to capture the specific features of the ML.
A clear identification and determination of appropriate roles and responsibilities in the MRM framework through the model entire lifecycle is another challenge for financial institutions. This is particularly important because of the increasing number and complexity of models used by multiple users of different areas of the financial institution. A specific point where the governance aspect of the MRM is crucial is the management of model changes. It is essential to have in place a process to assess the impact of the change and determine the review / approval steps appropriate for the materiality of the model change itself.
Inefficient reporting and fragmented systems represent another important challenge. The number of models is rising dramatically—10 to 25 percent annually at large institutions—as financial institutions utilize models for an ever-widening scope of decision making as ML11. As a consequence, risk managers must spend an increasing amount of time on non-strategic activities as data management, cleansing, and reconciliation.
