According to Cognizant (2015), model risk refers to the risk of loss that arises from various decisions that are made based on misuse or incorrect model reports and output. Model risk can be minimized through proper training in order to ensure accuracy in the development of models, their implementation as well as their use. Having proper control and governance related to these key stages of various models can reduce the human component in model risk by a large extent (Cognizant, 2015). This paper examines the human component in model risk in details.

You're lucky! Use promo "samples20"
and get a custom paper on
"Model Risk and the Human Error"
with 20% discount!
Order Now

Examples of common models in many organizations are the financial models, which are tools that are usually used to quantify risks or prices. They involve the application of mathematical relations, which advocate for the use of judgement and data in computing indications and parameters on how they should be applied in practical scenarios (Morini, 2011). Therefore, a model involves all these aspects. In view of this, it is a grave mistake for the people that use models to view them simply as mathematical functions without paying keen attention to how they are applied (Morini, 2011). Doing so introduces the human component in model risk. This is because model risks can arise depending on the way they are used. In some instances, the users might have a good model but end up feeding it with the wrong data or wrong parameters. This exposes a firm to subjective decision making that is based on the wrong output (Cognizant, 2015).

In most cases when people mention the model risk, they usually refer to the risk of having incorrect or flawed models (Borodovsky & Lore, 2000). It is worth noting that in today’s world, the traders normally rely on mathematical models, which involve advanced mathematics and complex equations. In view of this, model risk may arise when the models are being developed whereby the developers can make a mistake when setting the equations (Borodovsky & Lore, 2000). They can also make wrong assumptions especially when they are setting up the underlying asset price processes. To illustrate this further, one should consider a situation where the model designers make the wrong assumption about the interest rates (Borodovsky & Lore, 2000). This might make them base their model on a fixed and flat term structure instead of basing it on the actual term structure, which could be steep and unstable. Such human errors can lead to a situation where the model would be giving information that is inaccurate and could expose a firm to several risks (Borodovsky & Lore, 2000).

Nevertheless, model risk which is caused by the human error can be mitigated by proper training on the people who develop and use various models in an organization. A consultant should work with the risk management department in order to minimize the model risks. A consultant should also work with the human resource department in order to ensure that the people who are using various models are well trained and educated (Borodovsky & Lore, 2000). Since the human component of the model risk also depends on the way a model is used, the consultant should work with the human resource department to ensure that all business managers acquire the necessary training that would make them understand the mathematics and assumptions behind risk analysis. This training should also be conducted on employees of certain key departments who often interact with various models. Examples of such departments include corporate governance units, internal audit department, and the accounting department (Borodovsky & Lore, 2000).

There is always a human component in model risk whether in the development stage or the implementation stage. Due to this, it is possible for an organization to have correct models which are based on right assumptions but end up being applied in a wrong way. In order to reduce human error in model risk, a consultant should work with the relevant departments to ensure that employees are well trained in order to avoid human mistakes at development, implementation, and usage stages.