Schweser says “when the var is estimated using the historical simulation method or monte Carlos simulation we have all the values equal to or less than the var loss so it is straightforward to take the average of these to gey the cvar” how does this make sense? Cvar is the point after the var. How does taking the average of all the values that are either equal to or below the var point expected to get me Cvar? Confused
Var with given level of confidence simply tells you what is a minimum expected loss in tail. F. ex. with 95 % conf. level we might be sure that losses with no exceed particular VaR calculated amount in given period or there is (1-0,95) chance that minimum loss in given period would be this VaR amount. Thus, ordinary VaR does not say anything about further loss distribution if we reach that minimum loss in given period. To more precisely determine our loss exposure, we would like to know the amount of average loss in this tail if maximum loss given VaR would be reached within period. One of the methods is a conditional VaR which shows us an average loss exposure if VaR loss is exceeded within period. F. ex. There is a 5% chance that we would reach minimum loss of $ 15 M but if we exceed this loss of $15M is likely that average losses (over $15 M VaR threshold level) would be $20 M. This is expected loss distribution with 5 % chances to occur within given period.
1 Like
Conditional VaR is the average of all the losses in exceedance of VaR on the distribution curve.