We are traders

Performance, Policy, and Implementation of VaR

Provided a fully comprehensive analyses of policy, performance, and implementation of VaR methods and systems for a well known international investment bank.  The project covered all aspects of the VaR "loop" from: 

basic theoretical considerations of each flavour of VaR (why CVaR, HVaR, etc can provide "different" answers), 

model verification of theoretical formulations as well as well known "off-the-shelf" software packages.

data integration and effects (e.g. how choice of volatility surface generation may affect VaR), 

assessment of the complete VaR loop from real-time data » portfolio interface » VaR calculation » VaR audit

creation of manuals, materials and analytics to illustrate the effects of various assumptions in the usage and implementation of VaR such as:

- effects of sample length, for various weighted averaging methods for "calm" as well as "turbulent" market conditions

- effects of "indexing" VaR (such as CAPM, FX, correlation along interest rate curves etc).  For example, a CAPM based implementation of a particular portfolio was reported to have effectively twice the VaR as that of the non-CAPM based calculation under certain market conditions.

- back testing analytics to illustrate the performance of VaR for real portfolios and market conditions.  So, if the VaR measure is set at 95% confidence, do the P&L's actually lie below the 5% VaR-level 5% of the time;  if not than either the traders are being "over restricted" and so not making the required returns.  Or, alternatively, the firm is taking more risk than it had intended for the expected level of returns.  For example, a particular portfolio breached the 5% limit only 2.4% (i.e. less risk than VaR indicates) of the time over a 3-year (daily) VaR test, while another portfolio breached the 5% limit 7.5% of the time (i.e. more risk than VaR indicates). 

- effects of drift (trending) vs. volatility (e.g. VaR measures sometimes ignore drift, in the case that VaR measure is longer dated, such approached will mis-specify the risk exposure)  The Figure to the right illustrates a surface that can be used as both the measure of market conditions that require drift adjustment and the amount of the drift adjustment) such as the in the Figure below

 

If you are interested in obtaining research results on this issue please Request More Information and please feel free to indicate a few specifics of interest to you.

 ARTConsulting/Research

Back to Consulting Sample/Current Page

___________________