v5 p245

“VAR failes to incorporate positive results into its risk profile, and as such, it arguably provides an incomplete picture of overall exposures”


I can’t wait till this is over.

It is looking only at the losses side of the equation. So the above statement is TRUE.

Normal distribution is symmetric, one could just as easily restate VAR as not gaining more than X amt within a given confidence level

but that is not the point of VAR.

  1. It is not the maximum loss that could be suffered.

  2. It is the minimum loss that could be suffered.

  3. only looks at the loss side of the distribution. Even in the Historical VAR - Look at the losses - and take the nth worst possible loss.

Restatement still does not solve the problem, does it.

if you are taking a 5% VaR - you are looking at the left tail, 5% (1.65*Std Dev).