seemingly simple - shortfall risk vs. semivariance

An analyst is considering various risk measures to apply to a bond portfolio. She requires a measure that accounts for the magnitude of the losses. Given this requirement, as she considers using the semivariance and/or shortfall risk, she would reject using: A) both shortfall risk and the semivariance. B) shortfall risk but not the semivariance. C) the semivariance but not shortfall risk.

b shortfall risk is a probability

A

A

Shortfall risk doesn’t account for the magnitude of losses, but I believe that semi-variance does. The drawback to semi-variance is a reduced sample size. B… Final Answer.

B

B

C

B is correct. Shortfall risk gives an indication of the probability of not achieving a minimum return. The semivariance gives a measure expressed in returns. can someone explain how semi variance gives a “measure that accounts for the magnitude of losses?”

Shortfall risk is a probability of not meeting some target. The question says the measure must account for the magnitude of losses, this doesn’t. Semivariance is the standard deviation, but only for downside deviations. You compute get an estimate of magnitude of losses attached to probabilities with semivariance (providing you are willing to ignore fat tails), so it should work. That would mean answer B: reject using shortfall risk, but go ahead and use semivariance.

mumukada Wrote: ------------------------------------------------------- > can someone explain how semi variance gives a > “measure that accounts for the magnitude of > losses?” Looks like our posts overlapped… With semivariance, you construct confidence intervals - e.g. 90% sure you’re going to get more than this amount of return (or less than this amount of loss), implying that there is a 10% probability of having a larger loss than X. Usually this number is expressed as a percentage loss, but you can turn it into a dollar figure by multiplying by AUM. Value-at-risk is probably the most common methodology this fits into, although I’m pretty sure you could use semivariance in other ways too.

bchadwick Wrote: ------------------------------------------------------- > mumukada Wrote: > -------------------------------------------------- > ----- > > can someone explain how semi variance gives a > > “measure that accounts for the magnitude of > > losses?” > > > Looks like our posts overlapped… > > > With semivariance, you construct confidence > intervals - e.g. 90% sure you’re going to get more > than this amount of return (or less than this > amount of loss), implying that there is a 10% > probability of having a larger loss than X. > Usually this number is expressed as a percentage > loss, but you can turn it into a dollar figure by > multiplying by AUM. > > Value-at-risk is probably the most common > methodology this fits into, although I’m pretty > sure you could use semivariance in other ways too. What are you doing in the Level III forum; I thought you were already a Charterholder and didn’t stray from the General Discussion? Decide to help out the downtrodden class of 2009 or just bored with GD?

Both.

don’t shoo him away bankin! we need him :slight_smile: thanks bchad…that helps… now all I have to pray for is that at least one question on this shows up!!