VAR

CFA Level III Derivatives
#1

Can we aggregate VAR of different desks in the portfolio?

#2

The addition will be over estimate of actual overall VAR because you are not considering correlation between VARs.

#3

It does not aggregate easily into total var.

#4

Higher correlation, you sum almost all the VARs

Lower correlation, the total is less than the sum of all.

With low correlation, it’s unlikely that all classes (or portfolios) will be at risk at the same time.

#5

That means that you can do it. Thanks.

#6

Agreed that you can’t just sum up every desk’s VAR to come up with total VAR but what’s the correct formula for doing so?

#7

You measure VAR using all of the positions at the firm, not sure an aggregate of each positions VAR.

#8

I’d imagine return is volume weighted but are we supposed to know how to calculate vol from multiple portfolio vol each with different correlation? can’t remember if that’s part of the cirriculum

#9

You calculate portfolio std deviation with the weighted average of the returns

#10

ah thanks, i forgot about that formula

#11

If exam wants you to use conservative approach, add them up!

If exam want you do be reasonable, consider diversifaction effect.

#12

Except that you can’t.

At least, not without knowing that their correlations are all +1.0.

#13

If all are equities, normally we will add them up. (Conservative approach) I believe if it’s the same asset class, you can do it. (Of course, I know it ignores the diversificaiton effect)

#14

You can do anything you like.

The question is whether or not the answer you get will be correct.

It won’t be.

#15

Schweser, R26, p60.

Advantage of VAR:

  • It aggregates all risk into one single, easy to understand number.
#16

but is that number right? is that number what it was meant to be? Those are the questions you need to answer and understand first. That is what S2000 is saying - you can do whatever you like - but more of then than not, you do not know if that is the right thing to do.

easy to understand, but often misinterpreted. If you aggregate different types of risk - your VAR number is likely to be bigger than if the correlations (correct number) were used. But that is something that is not clear or discussed. And even if you did that, arrived at a number - when you start to use that number to control the /monitor risk on your business - you might find out - you have overestimated the risk - so are being more conservative in running your business, or you might go the other way - underestimate your risk - and cease to exist completely.

so go figure.

#17

Thanks cpk123, s2000. Got it.

#18

Limitation of VAR:

The difficulty of estimating standard deviation in very large portfolios. Standard deviation is generally computed from asset correlations and covariance. The required number of covariances for each pair of assets increases geometrically with the number of assets.