By Andrea Burgtorf, Head of Group Risk Operating Office, Erste Group.
Andrea, can you please tell the Risk Insights readers a little bit about yourself and your experiences with stress testing?
Over the last years, we have experienced a trend towards more quantitative and model-based methods to calculate risks; I personally have experienced this trend towards quantifying risks for more than 10 years. Particularly in stress testing, the methods have progressed significantly and the scope of stress testing has emerged from simple sensitivity stress tests to complex group wide stress tests with focus not only on economic scenarios but also increasingly on idiosyncratic scenarios and the combination of both. Moreover, current models should include both “financial” (i.e. Credit- and Market) risks as well as “non-financial” risks, formerly often just referred to as operational risk. In a wider context, this now also includes topics like cyber risk, conduct risk, reputational risk etc. Capturing all these different risk types is challenging – particularly with purely quantitative methods.
My experience also shows, that due to the change of job profiles for risk managers, i.e. more quantitative background required even for senior professionals; it is particularly important that senior management keeps both views on the topic, the quantitative as well as the professional expert judgment. An institutional memory as well as historic observations of past crises are very important in this context. No crisis will be alike. Therefore, models also have a bias towards mimicking the past: particularly due to the fact that often a long data history is observed. The professional risk manager combines quantitative results with qualitative experience and market know how to achieve the best possible outcome.
At the Stress Testing Europe Summit 2017, you spoke on your insight regarding ‘Reverse stress testing.’ Why do you believe this is key talking point currently in the industry?
Reverse Stress Testing gives you the opportunity to look at extreme events with the possibility to analyse severe outcomes for your institution. This is particularly important in order to quantify extreme losses and identify weaknesses in your portfolio, i.e. positions with high volatility, large loss potential or insufficient mitigation. It is mainly an intellectual exercise and should not be limited to the regulatory exercise but rather seen as a chance to improve portfolio quality, reduce position with high capital requirements and low diversification effect. Cost for mitigation of high risks should be analysed and ultimately have an impact on the consequences drawn from a proper reverse stress test exercise.
What are the key considerations that need to be made when constructing realistic scenarios for severe outcomes?
The challenge for reverse stress testing lies in the selection of still possible, but significantly severe scenarios in order to drive capital positions down below a minimum threshold, or, (which is even more important) impact the liquidity position of the bank in a way that a “going concern” assumption is not possible anymore. Often, an issue is the senior management acceptance of such a severe scenario which is most likely a combination of an economic stress and an idiosyncratic hit for the institution. Such a scenario has a very low probability and hence, there is a risk of it not being considered as a serious or realistic scenario. While there is good reason for this belief, I would still see the selection of a severe scenario; even under the assumption that it only has a probability of less than 1-2 % as an intellectual challenge; which should warrant a close look at the positions under such a scenario. Communication of the scenario as well as the background and reason for this type of stress test is important to give senior management like the board and supervisory board a good chance to understand the portfolio and its volatility even better, despite its low probability.
How do you see the stress testing industry evolving over the next 6-12 months, particularly with the current and upcoming political environment in mind?
Stress testing will continue to be an integral part of a sound risk management practice. Enhancing regulatory stress tests by own scenarios, which are suited to analyse the portfolio and its vulnerabilities best, require a sound knowledge of crisis patterns, contagion effects and an idea about the evolution of financial markets during and after a crisis. Only with this will we know how it is possible for the stress test design and selected scenario to suit each individual institution best.
Finally, there is only value to the institution, if you draw your own conclusions out of the results of the stress test, i.e. what is most important, including the identification and implementation of mitigating actions and a respective follow up after implementation is in place. Making an impact to the existing portfolio is the best purpose of a severe stress test.