STRESS TESTING-INSIGHTS ON RISK MANAGEMENT
PRMIA Toronto would like to invite you to attend the Global
Event Series on Stress Testing.
The recent financial crisis has highlighted a number of shortcomings
in risk managements, especially when liquidity dries up or
large tail events occur. Combining traditional risk metrics
with additional risk measurement tools such as stress tests
adds qualitative judgement to the existing quantitative framework.
Join us for an insightful discussion on stress testing with
industry experts. Attendees will also have the opportunity
to network and enjoy light refreshments in a relaxed environment.
Melissa Van Hees (Concordia Advisors)
Stress Testing: A practitioner's perspective
Financial instability is here to stay, therefore stress testing
will continue to be an important part of the risk managers'
tool kit. Stress tests in the past used to mostly focus independently
on each type of risk (market, counterparty, credit, liquidity,
operational) and were performed inside the risk function with
little dissemination to outside groups. The tests also tended
to rely too much on probabilistic approaches. Stress testing
must be comprehensive, integrated, and incorporate systemic
considerations. The presentation will touch on some newer
concepts such as reverse stress tests, varying time horizons,
and stressing the entire balance sheet. Infrastructure and
changes to existing governance frameworks will also be discussed.
Dr. Lance Smith (Imagine Software)
Stress testing has become the pre-eminent tool for portfolio
risk management, supplanting the traditional parametric VaR
calculation. As a consequence risk management has, in a way,
become more subjective in that there is no clear methodology
for designing these stress tests: what factors should be stressed,
by how much and with what probability? In addition, when stressing
a particular factor, how can we quantify the collateral effect
on the remaining factors? The parametric VaR calculation,
while deficient in some ways is not nearly as subjective in
that the inputs are usually computed from time series using
standard statistical algorithms. We seek a way to meld both
methods together in a way that maintains the intuitiveness
of stress testing and the statistical rigor of parametric
VaR. This idea can be extended to the nonlinear factor approach
to risk management, as well as to empirical distributions.
Event moderator: Grigoris Karakoulas (Infoagora)
5:15 pm Opening/ registration
5:30 pm 6:30 pm Speaker presentations
6:30 pm Questions and Answers, Networking
PRMIA Toronto would like to thank Imagine Software for sponsoring