Guide

Model risk

Paul Young Paul Young

Models are notoriously complex. Despite this, so much of modelling happens inside a ‘black box’ environment and the development team are often the only people who really understand the calculations applied. But getting it right is vital as the impact of a miscalculated model can be devastating.

Effective models require continual checks

Regardless of the complexity, you need to know what your models are doing. That means having the skilled resources to offer genuine challenge to the development team. They should check that the logic works, the inputs are reliable and the model is being used for its intended purpose.

While these sound like simple checks, they aren’t. Nor are they a one-off before making the model live. They should be performed regularly, reflecting the fact that models have a lifecycle of their own. As your business changes and evolves, so too will the inputs required for your model. They are time limited and need regular validation and re-calibration to remain accurate.

Making informed decisions

So why does any of this matter? It matters because your firm relies on models in order to demonstrate what would happen to your business in certain scenarios. The outcomes of the models inform a number of business decisions and, if that data is wrong, it can have a big impact on the firm.

Not only does it raise operational risk, but you could also face major financial consequences. In the past, firms have suffered eye-watering losses of up to $6.5 billion on their projections, partly due to mis-calibration of a model. Many firms couldn’t handle such crippling losses - do you want to put your business to the test?

Keeping up with regulatory requirements

For banks above a certain size, modelling is a regulatory requirement and they are currently facing increased regulatory scrutiny. Recent guidance from the European Central Bank (ECB), Prudential Regulatory Authority (PRA) and the US Federal Reserve outline similar views of best practice for managing model risk. So getting risk management right is important across the board.

Ensuring the best defence

Truly embedding any risk management framework requires vision. It needs creativity to seek out areas within the business where controls would be best applied. You should review your three lines of defence model and consider how each level can help to mitigate model risk. Once these controls are in place, it’s worth getting assurance over their effectiveness, as well as independent validation over your individual models.

For more information on model risk and how we can support your business, contact Paul Young.

Download our guide
Download PDF [ 521 kb ]