What is Model Risk?

What is model risk? When discussed in class, Professor Grace stated the summarized definition of "when a firm isn't looking at what's really going on due to them being stuck on a particular model, even if that model is not the correct one." Riccardo Rebonato comprised a definition as well. It simply states that model risk occurs after observing a set of prices for hedging instruments, and receiving a different result after using different yet similar models. Because of the mix-match of results, financial losses are incurred.

There are different types of model risk. 1)Wrong Model- This is when the wrong model is chosen to measure a particular project, returns, or a set of data. 2)Model Implementation- During the process of applying the chosen model, different complications can occur such as using wrong data, estimated values, or other technical errors. 3)Model Usage- This happens when a model is used improperly.

Why is model risk important? What affects does it have on a firm? Looking from its definition, model risk can cause many financial losses. For example, financial institutions rely greatly on models for their investment projects, investment returns, and other data. They use results from created models to determine their moves in the industry. Because most moves require great amount of resources (both money and time), it is critical that they models they used are correct and portray the correct future outcomes. If model risk is high, firms won't only lose financially, but could potentially lose their reputation.

Model risk is very important. Because of all the factors that it affects, firms allocate a lot of money towards model risk management. Once firms are better at choosing the correct models and implementing at the appropriate time, they will decrease their model risk and keep potential monetary and reputable losses low.

(Sources for this post are from "Theory and Practice of Model Risk Management" by Riccardo Rebonato and Wikipedia: "Model Risk".)