Please see attached for question 39. In the answer, then they calculate VaR they subtract the mean from the st. dev. When they calculate the liquidity adjustment or spread, they add the mean to the st. dev. Is this correct? Any insight much appreciated.
They have answered correctly here ... the subtract in the VaR is due to the expected return, which offsets the potential loss
however, in regard to the spread, we just want the "worst probable" spread so it is added
the +/- can give confusion but they have basically followed Dowd here, which is what i recommend to avoid +/- confusions. What i mean is, our best (absolute) VaR is given by:
VaR (%) = -return*T + volatilty*deviate*SQRT(T)
i.e., that's what they did in the answer ... this produces (except for long long VaR) a postive VaR which represents the potential loss
then the exogenous liquidity adjustment adds to this positive (representing greater potential loss)
LVAR adjustment = 1/2*(mean spread + volatility of spread*One-tailed deviate)
i.e., add because we are concerned about the spread widening against us
further note that GARP has *correctly* used the one-tailed deviate. They used 2.33 instead of 1.96!