Supervising Model Risk
It has been stated more than once that Model Risk had a massive role to play in the Global Financial Crisis and subsequently; there has been substantial research into the role risk modelling played to lead stakeholders to make the wrong decisions when facing uncertainty.
"Risk management, even if flawlessly executed, does not guarantee that big losses will not occur. Big losses can occur because of business decisions and bad luck. Even so, the events of 2007 and 2008 have highlighted serious deficiencies in risk models."
Risk Management Lessons from the Credit Crisis | Philippe Jorion [LINK]
Further investigations from regulators including the Bank for International Settlements has found that variations in the way banks risk-weight their assets, even for identically designed hypothetical portfolios differs considerably.
Model risk is an elusive and obscure category of risk that is coming under further scrutiny from regulatory bodies and most financial institutions need to put something in place to address this threat.
The risk needs to be uniquely defined too and Wikipedia has this label for it; "The risk of loss resulting from using models to make decisions" [LINK]. That seems simple enough doesn't it but the reality of the matter is far from simple.
Additionally, there are lots of paradoxes that tend to playout that make model risk complex to understand in itself and one can't always lay blame on the model but the user of it. We have to be very real with model risk and risk reports in general. When you see a small probabilistic measure of anything, very small, tiny, perhaps even improbable; there is still a small slither of a chance that the unwanted situation will eventualise.
This morning I stumbled across this presentation from the system provider SAS which I have to say is rather grand and definitely worth taking a look at. I recommend risk managers peruse the slides within because there are some very useful references on model risk and the direction regulators are likely to take with respect to this risk management concern.
Impact of supervision on Model Risk | SAS [LINK]
In Causal Capital, we are looking to develop a complete framework that will assist auditors assess the capability of their relevant institutions to address model risk and all faring well we should be publishing our white paper in the next few months.