Backtesting and stress testing
In finance, a stress test could be viewed as an analysis or simulation designed to determine the ability of a given financial instrument, such as a VaR to deal with an economic crisis. Since the first method to estimate a VaR is based on the assumption that stock returns following a normal distribution, its accuracy depends how far, in the real world, stock returns deviate from this assumption. A key component to the implementation of model-based risk management is model validation. That is, we need some way to determine whether the model chosen is accurate and performs consistently. This step is quite important both to firms and their regulators. According to Lopez (2000), we have the following table:
Name |
Objectives |
Methods |
---|---|---|
Backtesting |
Compare observed outcomes with a model's expected output |
Forecast evaluation established empirical issue with a large academic literature |
Stress testing |
Examples a model's expected outcomes under extreme conditions... |