I tried, tried and tried again to make some sense out of the concept of VAR but my brain refuses to intake. Then again risk management in particular and statistics in general is not my strongest suit. What is the likelihood of VAR being tested in the exam? I remember reading somewhere that it was not tested at all in 2009 exam. Is this correct? retakers?
VAR is quite an easy concept, honestly. It’s the “chance” (probability) that your assets will fall below a certain threshold value. The major shortfall is that it can’t tell you the magnitude of your loss, only that there’s a _% chance that you’ll lose more than $____ in a given time period. Another is that it assumes a normal distribution, which ain’t too realistic. That’s really all you need to know, as far as I understand it.
Here’s a simple example: You have a portfolio of $100MM. Yearly standard deviation of 5. Now you’re just dealing with a normal probability distribution. So, since the odds of a number falling outside three standard deviations are ~1%, we can say that at a 1% confidence level, your VaR is 15 (3X5). At a 5% (I think the z-score is 1.96, but not positive), your VaR would be 9.8. Etc., etc.
Thanks Skillionair: This is really encouraging and helpful. I guess I lack basis foundation of probability distribution. For example, you said “since the odds of a number falling outside three standard deviations are ~1%,…”. Where do you get that? What is z-score (is it a standard table?) and how do you get 1.96 in that? etc. Also, do you have to learn calculations for all three methods of VAR (analytical, historical and Monte Carlo) or just the advantages and disadvantage? Schweser makes it seem like you have know calculation of only historical method. Again, I really appreciate you help.
There’s a thread (a few threads down) that deals more in-depth with historical VaR vs. MC VaR that’d probably be good for you to read. And as far as the standard deviation/probability stuff, that’s all just stuff that (I think) I remember from earlier levels, using it, etc. I’m not really the guy to be asking regarding probabilities, confidence intervals, and z-scores - I was just trying to walk you through a quickie example to show you the concept.
skillionaire Wrote: ------------------------------------------------------- > VAR is quite an easy concept, honestly. > > It’s the “chance” (probability) that your assets > will fall below a certain threshold value. > > The major shortfall is that it can’t tell you the > magnitude of your loss, only that there’s a _% > chance that you’ll lose more than $____ in a given > time period. > > Another is that it assumes a normal distribution, > which ain’t too realistic. > > That’s really all you need to know, as far as I > understand it. Sorry, but you are missing a ton, analyitcal Var (variance-convariance), historical VAR, and Monte Carlo Var. Not all VAR needs to have normal distribution, that’s only a dis advantage for analytical VAR Extension of Var also includes, IVAR, CF VAR, Tail Risk Var, etc.
whystudy Wrote: ------------------------------------------------------- > skillionaire Wrote: > -------------------------------------------------- > ----- > > VAR is quite an easy concept, honestly. > > > > It’s the “chance” (probability) that your > assets > > will fall below a certain threshold value. > > > > The major shortfall is that it can’t tell you > the > > magnitude of your loss, only that there’s a _% > > chance that you’ll lose more than $____ in a > given > > time period. > > > > Another is that it assumes a normal > distribution, > > which ain’t too realistic. > > > > That’s really all you need to know, as far as I > > understand it. > > Sorry, but you are missing a ton, analyitcal Var > (variance-convariance), historical VAR, and Monte > Carlo Var. > > Not all VAR needs to have normal distribution, > that’s only a dis advantage for analytical VAR > > Extension of Var also includes, IVAR, CF VAR, Tail > Risk Var, etc. Yeah, I wasn’t trying to give a comprehensive tutorial on VaR and its extensions (and actually pointed the OP in the direction of another thread when asked to do so) - I was just trying to help the OP “make some sense out of the concept” of VaR.
Thanks guys. I spent 3 hours today reading everything about VAR in Schweser and CFAI and I think now I have good understanding of the concept. It wouldn’t have happened without your encouragement. Thanks again.