We have seen the importance of risk models and risk management in general for the successful functioning of an organization. However, the risk models that are in place today have certain shortcomings.
Most models work with a one-year time horizon. However, a one-year time horizon may not be sufficient in all situations, and the model may not be designed to scale for a longer horizon. The extended time horizons become more valuable in the times of crisis. For example, in the times of crisis the markets may become very illiquid, and a position that was expected to be held for just a day or two may have to be held for months now. Second, the firms using daily VAR may have to accompany it with other measures for longer period because during crisis, the everyday losses may exceed daily VaR continuously.
Another problem is that the existing statistical risk models assume risks to be related just to the firm (Exogenous). They ignore risk concentrations across the financial institutions. For example, if a financial firm has exposure to sub-prime sector, the model will ignore that many other financial institutions also have exposure to the same sector which multiplies the risk.
The risk models may not also account for possibility of situations such as predatory trading. In such a situation the trades made by other traders affect your positions and compound your losses.
Since these risk models use historical data, it’s difficult for them to incorporate such complex situations. The risk management can think of complementing these risk models with other tools such as scenario analysis to deal with the crisis situations.
What other shortcomings do you think the existing risk models have?
This article is based on the paper “Risk Management Failures: What are They and When Do They Happen?” by Rene M. Slutz, which is a part of the FRM syllabus.