Stress testing can be defined as a tool, which managers of firms use to simulate a possible, but adverse event and project, the extent of occurrence of which would affect the income of the firm (Kupiec, 1998). The tool usually examines the probable effects on a given firm’s financial status of changes in factors of risk that correspond to plausible but exceptional events (Schachter, 1998). The tool has various distinct characteristics that set it apart from other risk management tools. They are as following:
- It is non–statistical and measures potential loss through considering a set of worst case scenarios;
- Each of the scenarios presented defines different and distinct risk factors that are bundled together, and the extent, to which each of the sets of factors is likely to move under a given scenario;
- Each scenario’s impact is determined through the application of the scenario to the accrual and/or trading position of the firm, and then calculating the loss in each scenario. The scenario that exhibits the biggest loss is referred to as the stress-loss.
The use of the tool is imperative to the management of firms due to the following reasons:
- It helps them to address the key risk issues, determining the impact of P&L at a given stress scenario, and determining the worst-case scenario, and the losses associated with the same;
- It complements VaR, since the latter does not inform on the exposure direction, does not quantify the probable size of losses in cases, where the actual loss is greater than VaR, and it fails to take into account the stress events.
Effective Stress Testing Techniques.
There are various stress testing technique methodologies, which application has proven to be relatively effective. The three main ones are as following:
• Sensitivity analysis;
• Hypothetical scenarios;
• Historical scenarios.
However, there are other methodologies as well that are in limited use, and which include the Extreme value theory, which uses history of stress events that occurred in the past in determination of the severity and frequency of occurrence of stress events in future. Another uncommon methodology is Stressed VaR.
Implementing a stress testing program for market risk
In the implementation of a stress testing program, it is important to make various informed decisions before the actual application of the chosen program (Dowd, 1999). For the market risk, historical scenario analysis is especially important. In this analysis, it is of the extreme importance that various aspects are considered. Some of these are as stated below:
- The most appropriate scenarios. The implementers should make sure that they identify the relevant scenarios that have a real possibility of occurring to the business;
- They should also ensure that they pick an appropriate period of the scenario that is representative of the whole scenario, if and when the scenario period is too large to be considered as a whole (Berkowitz, 1999);
- Before the onset of the initial program, the implementers should decide whether to go for relative (use percentages), or absolute risk factor changes;
- It is also important to identify the relationship between the regime changes and possible risk factors;
- The systems should be in a position of supporting current and future events in terms of flexibility and robustness;
- The correct personnel should be chosen for the job. The modeler should be highly skilled, as it is only then that they can identify the most suitable scenarios;
- The costs that come with computation should be put in consideration to avoid hiccups during the initial implementation.
In the implementation, there are some critical technical issues that the management needs to consider. Some of these are:
- The pricing models - whether Full Revaluation, Grid-based, Full Revaluation, or Taylor- series will be applied;
- The holding period. It is imperative to make the right assumption, whether the period is going to be a week, a month, or a year;
- How frequently the stress testing should be made;
- The correlation assumptions.
For the program to be highly effective various measures should be undertaken (Arya, 2008). Such measures include the following:
- Review of the stress scenarios on the regular basis;
- Frequent, daily or weekly, running of the report of the stress test results;
- Involvement of senior management in the program.
The Stress test programs should be in line with the financial institutions regulator’s expectations. Examples of such expectations in many jurisdictions are as following:
- The scenarios used should be mandated by banking supervisors. However, the use of some scenarios is not mandated;
- Proper documentation of the scenarios should be held;
- The tests have to be both qualitative and quantitative;
- The tests should incorporate market disturbances’ liquidity aspects in addition to the market risks;
- The tests should be periodic and has to be performed at the regular intervals.
It is important to note that though stress testing gives a firm information on what is most likely to happen to the said firm under the certain scenarios, it does not predict the likelihood of the actual occurrence of the scenario. Stress testing programs should, therefore, be undertaken hand in hand with supplementary programs that accurately predict the likelihood of occurrence of each scenario in the program in order to enable the firm’s management to know which programs to act upon. The proper implementation of a stress test program will ensure that a financial institution is able to prevent crippling losses (Aragones, Blanco & Dowd, 2001).