Model Risk

Model risk is the risk of loss or poor decisions arising from errors in, or misuse of, a financial model. As banks and businesses rely ever more heavily on models to price assets, measure risk and make decisions, the danger of those models being wrong — or being used wrongly — has become a serious risk in its own right. This guide explains what model risk is, where it comes from, how it's managed, and why it matters — in plain language. It connects closely to the validation of models and is a core topic in risk qualifications like the FRM.

What is model risk?

A model is a simplified, mathematical representation of reality used to help make decisions — pricing a derivative, estimating credit losses, or calculating capital requirements. Model risk is the potential for adverse consequences when that model produces inaccurate outputs or is applied inappropriately, and those outputs are then relied upon. Because models inevitably simplify a complex world, every model is "wrong" to some degree; model risk is about the danger that those imperfections, or their misuse, lead to real losses or flawed decisions. As the statistician George Box famously put it, "all models are wrong, but some are useful" — model risk is what happens when a useful model is mistaken for a perfect one.

Where model risk comes from

Model risk generally arises from two broad sources:

• Errors in the model itself. The model may be built on flawed assumptions, contain mathematical or coding mistakes, use poor-quality or unrepresentative data, or rest on a theory that doesn't hold in the real world. A classic example is a model that assumes asset returns are normally distributed and so underestimates the chance of extreme events.
• Misuse of the model. Even a sound model can cause harm if it's applied to the wrong situation, fed inappropriate inputs, or its outputs are misunderstood — for instance, using a model calibrated for normal conditions during a market crisis, or trusting a precise-looking number without grasping its underlying assumptions.

Why model risk matters

The dangers of model risk have been demonstrated repeatedly. The 2008 financial crisis is the most prominent example: models used to value complex mortgage-backed securities and to rate them relied on assumptions — such as that house prices wouldn't fall sharply across the whole country at once — that proved disastrously wrong, contributing to vast losses. Earlier, the collapse of the hedge fund Long-Term Capital Management in 1998 was driven in part by sophisticated models that failed when markets behaved in ways the models deemed almost impossible. The common thread is over-reliance on models whose limitations were underappreciated — treating a model's output as reality rather than as an estimate built on assumptions.

How model risk is managed

Financial institutions manage model risk through a discipline often called model risk management, which typically includes:

• Independent validation. Having a separate team rigorously test and challenge a model before and after it's used — checking its assumptions, data, mathematics and performance.
• Governance. Maintaining an inventory of models, clear ownership, documentation, and approval processes so models are understood and controlled.
• Ongoing monitoring. Regularly back-testing a model's predictions against actual outcomes and reviewing it as conditions change.
• Understanding limitations. Crucially, ensuring those who use a model's outputs understand its assumptions and where it can break down, so it's applied with appropriate caution.

Regulators now expect banks to manage model risk formally, reflecting how central models have become to finance.

Why it matters for finance professionals

As finance grows more quantitative, model risk becomes ever more important. Understanding that models are tools with limitations — not oracles — is essential for anyone who builds, validates or relies on them. The ability to question a model's assumptions and recognise where it might fail is a hallmark of sound risk management, and a regularly examined topic in professional qualifications.

Frequently asked questions

What is model risk?

The risk of loss or poor decisions resulting from errors in a financial model or from using it inappropriately, when those flawed outputs are relied upon.

What causes model risk?

Two main sources: errors in the model (flawed assumptions, mistakes, poor data) and misuse of the model (applying it to the wrong situation or misunderstanding its outputs).

Can you give an example of model risk?

The 2008 crisis, where models valuing mortgage-backed securities assumed house prices wouldn't fall sharply nationwide, and the 1998 collapse of Long-Term Capital Management, whose models failed in extreme markets.

How is model risk managed?

Through independent model validation, strong governance and documentation, ongoing back-testing and monitoring, and ensuring users understand each model's assumptions and limitations.

Build your risk skills with Learnsignal

Model risk sits at the heart of modern, quantitative risk management. Learnsignal's tutor-led courses, including the FRM, develop the risk understanding that topics like this build on — with clear teaching that stresses both how models work and where they fail.