LO 5.3: Com pare VaR, expected shortfall, and other relevant risk measures.

LO 5.3: Com pare VaR, expected shortfall, and other relevant risk measures.
VaR estimates the maximum loss that can occur given a specified level of confidence. VaR is a useful measure of risk since it is easy to compute and readily applicable. However, it does not consider losses beyond the VaR confidence level (i.e., the threshold level). In other words, VaR does not consider the severity of losses in the tail of the returns distribution. An additional disadvantage of VaR is that it is not subadditive, meaning that the VaR of a combined portfolio can be greater than the sum of the VaRs of each asset within the portfolio.
2018 Kaplan, Inc.
Page 57
Topic 5 Cross Reference to GARP Assigned Reading – Basel Committee on Banking Supervision
An alternative risk measure, frequently used by financial institutions, is expected shortfall. Expected shortfall is more complex and computationally intensive than VaR, however, it does correct for some of the drawbacks of VaR. Namely, it is able to account for the magnitude of losses beyond the VaR threshold and it is always subadditive. In addition, the application of expected shortfall will mitigate the impact that a specific confidence level choice will have on risk management decisions.
Spectral risk measures generalize expected shortfall and consider an investment managers aversion to risk. These measures have select advantages over expected shortfall by including better smoothness properties when weighting observations as well as the ability to modify a risk measure to reflect an investors specific risk aversion. Aside from the special case of expected shortfall, other spectral risk measures are rarely used in practice.
S t r e s s T e s t i n g
It is important to incorporate stress testing into risk models by selecting various stress scenarios. Three primary applications of stress testing exercises are as follows:
1. Historical scenarios, which examine previous market data.
2. Predefined scenarios, which attempt to assess the impact on profit/loss of adverse
changes in a predetermined set of risk factors.
3. Mechanical-search stress tests, which use automated routines to cover possible changes
in risk factors.
In stress testing, it is important to stress the correlation matrix. However, an unreasonable assumption related to stress testing is that market shocks occur instantly and that traders cannot re-hedge or adjust their positions.
When VaR is computed and analyzed, it is generally under more normalized market conditions, so it may not be accurate in a more stressful environment. A stressed VaR approach, which attempts to account for a significantly financial stressed period, has not been thoroughly tested or analyzed. Thus, VaR could lead to inaccurate risk assessment under market stresses.
Professor’s Note: In Book 3, we w ill explain the calculation o f stressed VaR.
I n t e g r a t e d R i s k M e a s u r e m e n t