MODEL RISK

Model risk is the risk associated with failure of some model used to calculate the value of the stock. In certain case the model totally fails to observe some risk associated in practical situation, due to this in some case the value calculated by this model fails miserably. It is a subset of operational risk. Traders who use these models are not supposed to know the risk associated with this mode. Traders often do not know the limitation and assumption associated with the models, and hence unaware of the risk associated with it. In modern days the computer are used to make investment decision. These computers use the model to find out the best stock. The volume on investment made using these computer models are quite high, which in turn makes the investment highly leveraged. A small error in the model leads to a huge loss.

There are basically three types of model risk they are,

• Wrong selection of model in this error the investor selects the model which does not suits the investment type, this will lead to a huge variation in the calculation part.
• Model implementation error is an error in which the even though the selection of the model is made correctly the model is not implemented correctly. There might be some error while implementing the model or some technical errors while implementing the model.
• Model usage error occurs when the model which is used inappropriately. The error in this type includes wrong date entry, risk related to implementation and the finally the error associated to the calibration of the model.

The error associated with the model cannot be completely reduced however it can be reduced drastically by adopting the following techniques,

• The model when being developed must include all the key elements. It must include all the risk associated with the investment. Model must be as simple as possible so that further enhancement can be made easily.
• The implementation risk must be averted. It must be done with utmost sincerity while implementing the model.
• Model before being used commercially must undergo all the stress test. It must be checked for correction to the extreme condition that it might face in the practical scenario.

Thus model risk must be made as much negligible as possible in order to make correct decision using the models. The model risk cannot be removed completely however can be reduced drastically using the appropriate precautions.