E[CAPM] & Market VaR Zero Mean Assumption

[mikey10011]

On slide QUANT A slide 21 the time series to calculate daily VaR was assumed to have an expected return of zero.

Question: Given that under CAPM the expected return E[r] is nonzero (= riskfree rate + beta x market risk premium), how do I reconcile the two?

 

[David Harper CFA FRM]

Mikey,

I'll assume you are reading not listening I don't blame you....The CAPM is right. Slide 21 refers to the TWO SIMPLIFYING ASSUMPTIONS we make (following Hull) including: assume mean return = 0. It is a "rounding off," that is technically INCORRECT but conventionally (by practice) okay b/c the daily mean return will be small; e.g., +20% annual is only (1/252) 8 basis points. It makes the variance easy: squared returns = variance. That's all. It is (even more) correct to include the mean return.

Pleaes note: we don't take this shortcut for longer intervals (monthly periods, annual periods) where the mean return becomes materially non-zero.

David