Model Risk Management

When it comes to managing risk, financial institutions have long relied on models. Models are used to calculate the probability of default for a loan, the value of a derivative, or the riskiness of a portfolio. But as the financial crisis of 2007-2008 showed, models can be flawed. In some cases, the models underestimated risk. In others, they failed to account for correlations among assets that became evident only in a crisis.

In the aftermath of the crisis, regulators around the world began to pay closer attention to model risk—the risk that a model is incorrect or misused. They issued new rules and guidance on model risk management, and they began to scrutinize models more closely. Today, model risk management is a top priority for financial institutions. They are investing heavily in model governance, validation, and documentation. They are also paying more attention to model risk when making decisions about product approvals, limit setting, and stress testing.

Despite these efforts, model risk remains a challenge. One reason is that models are constantly changing. Financial institutions must continually update their models in response to changes in regulations, market conditions, and their own business strategies. Another challenge is that model risk management requires coordination among multiple functions within a financial institution, including risk management, finance, accounting, and technology. Each function has its own objectives, priorities, and culture. And each function uses different models with different levels of complexity.

The coordination challenge is compounded by the fact that model risk management is a relatively new field. Standard approaches to model risk management are still emerging, and financial institutions are still learning what works best for them. Finally, model risk management is an ever-evolving field.

As the financial industry becomes more complex and interconnected, the risks posed by flawed or misused models are likely to increase. Financial institutions will need to continually adapt their model risk management practices to stay ahead of the curve