Risk Management

WIth increasing prevalence of machine learning and big data analytics, institutions  increasingly rely on data driven, quantitative analysis and models in various aspects of their decision making.

The expanding use of models in all aspects of decision making have proven to improve business decisions, but the benefits of models come with costs – both direct and indirect. The direct cost is to develop and implement models properly. Then there are potential indirect costs of relying on models, such as the possible adverse consequences (including financial loss) of decisions based on models that are incorrect or misused. Those consequences should be addressed by active management of model risk.

“Models are simplified representations of real-world relationships among observed characteristics, values, and events. This simplification is not only inevitable due to the inherent complexity of those relationships, but also is intentional  to focus attention on particular aspects that are most important for a given model application.

Models are never perfect, and the appropriate metrics of quality, and the effort that should be put into improving quality, depends on the situation. Hence in all situations, it is important to fully understand a model’s capabilities and limitations given its simplifications and assumptions.

Model Risk Management

Effective model risk management represents a wholistic approach incorporating all model stakeholders and is much more than periodic model validation exercise. It requires Institutions to have a model risk policy regulating the definition and scope of model risk management, model inventory, model approval process, model validation and management of model weaknesses.  

Involves performing back testing on lower portfolio levels for a long period of time, dividing back testing results into risk drivers. This is done for different quadrants based on dirty P&L along with clean P&L.

Risk measure consistency tests investigate relationship between stress tests and VaR as well as stressed VaR and VaR and comparing different VaR percentiles.

Understand the pricing model weaknesses and their implication for VaR, verify consitency of P&L and VaR calculation processes and analyze inconsitencies.

Conduct alternative VaR calculation on different pricing model and sensitivities. Compare alternative VaR models: covariance matrix, simulation etc.

Model Validation

A regular assessment of models should be conducted to create a model weakness inventory. Together with a complexity and materiality scoring of the model portfolios, a risk/materiality matrix is defined that enables identification of portfolios for subsequent in-depth validation. Results of model validation exercise are presented in a validation report carrying list of issues and validation results. 

The application of risk classifications to a company’s models is still not a perfect science. As such, the best practice is still developing. Typically, a risk classification would consider a model’s materiality and complexity. The use of the model is also usually considered, for instance, a model for regulatory reporting will typically have specific requirements.

We work with the client team to identify the appropriate risk scoring methodology taking into account all factors including their materiality, complexity and regulatory scruitiny. 

Validation is a key step towards approval to use an internal model. If properly designed and implemented, the validation process has the potential to enable continuous improvement and enhancement of the internal model and, as a consequence, improving the company’s understanding and management of risks 

The result of the validation process are produced in a report summarizing the scope of the validation, the governance, the processes and tools applied; confirming the model strengths and identifying the model weaknesses and limitations.

The report focuses on the key messages which also provide sufficient detail to give the Board and regulators the confidence that the conclusions were credible and robust. In order to support this process, we also provide education to the Board regarding internal model validation.

Our Offerings

Model Risk Governance

A regular assessment of models is conducted and a model weakness inventory is created. Together with a complexity and materiality scoring of the model portfolios, a risk/materiality matrix is defined that enables identification of portfolios for subsequent in-depth validation.

Model Risk Governance

A regular assessment of models is conducted and a model weakness inventory is created. Together with a complexity and materiality scoring of the model portfolios, a risk/materiality matrix is defined that enables identification of portfolios for subsequent in-depth validation.
LEARN MORE

Model Validation

A regular assessment of models is conducted and a model weakness inventory is created. Together with a complexity and materiality scoring of the model portfolios, a risk/materiality matrix is defined that enables identification of portfolios for subsequent in-depth validation.

Model Validation

A regular assessment of models is conducted and a model weakness inventory is created. Together with a complexity and materiality scoring of the model portfolios, a risk/materiality matrix is defined that enables identification of portfolios for subsequent in-depth validation.
LEARN MORE
You need more information

Frequently Asked Questions

You can outsource any back office or non-core business function, as we provide a wide array of services, some of which are: Call Center Services, Data Entry Services, Software Development Services, Mortgage Services 

Looking For Something More Specific?

Write to us at contactus@biznessanalytics.com or