Answer-first summary for fast verification
Answer: Models should produce a range of values for the bank's loss in a stress scenario, rather than a point estimate.
**Explanation:** C is correct. It is often more appropriate for stress-testing models to provide a range of possible losses rather than a single estimate. This approach acknowledges the inherent uncertainty in stress scenarios and provides a more realistic view of potential outcomes. A is incorrect. While models such as VaR models can be validated using back-testing, there is no similar way of validating the output from a stress test because stress testing deals with rare events. In this way, the validation of stress-testing models is more difficult than the validation of other models. B is incorrect. When there are doubts about the best model to use in a certain situation, the results from several models can be compared and the totality of stress-testing reports can make the attendant uncertainties clear. D is incorrect. Models describing the relationship between variables in normal market conditions (as a model using correlations averaged over a business cycle would) may not describe how they behave in stressed market conditions. Correlations tend to increase in stressed market conditions.
Author: LeetQuiz .
Ultimate access to all questions.
No comments yet.
A newly hired model risk analyst at a bank is validating a model that was recently developed for use in stress tests conducted by the bank. The analyst first consults a senior colleague regarding best practices for validating stress testing models. Which of the following statements would the analyst's colleague be correct to make?
A
Stress test models are typically back-tested against the bank's historical data to ensure that the model is specified correctly.
B
Each stress test should use a single model so that the relationship between the scenario variables and the output is clear and unambiguous.
C
Models should produce a range of values for the bank's loss in a stress scenario, rather than a point estimate.
D
Relationships between variables in a model used for stress testing should typically be represented using the averages of their correlations over one full economic cycle.