Just a quick Q on calculating VaR For the Analytical / Variance-Covariance method, the examples in the book all use the return and standard deviation as given (or as calculated using the usual portfolio return / standard deviation equations). There is a sentence or two that discusses converting distributions to standard normal distributions (subtract mean, divide by standard deviation). Questions: 1) I am not entirely sure when we would have to do this. Can someone explain why/when we would convert to a normal distribution? 2) Also, can someone explain what figure from which you subtract the mean and divide by the standard deviation? i.e. How is the conversion used? I am fine with the rest of the analytical VaR material. It is really just that one little section on conversion to a standard normal distribution.
[X-Mu]/sigma -> normalize each variable… is what I understood. x = each data point you are looking at. mu = average of all data points sigma = std. deviation of those data points. what you do to get at the limits of a confidence interval. may be wrong, and way off base
Ok, so if this is normalizing each variable (e.g. each monthly return), then we won’t be doing this on the exam, correct?
boost Wrote: ------------------------------------------------------- > Just a quick Q on calculating VaR > > For the Analytical / Variance-Covariance method, > the examples in the book all use the return and > standard deviation as given (or as calculated > using the usual portfolio return / standard > deviation equations). > > There is a sentence or two that discusses > converting distributions to standard normal > distributions (subtract mean, divide by standard > deviation). > > Questions: > 1) I am not entirely sure when we would have to do > this. Can someone explain why/when we would > convert to a normal distribution? > > 2) Also, can someone explain what figure from > which you subtract the mean and divide by the > standard deviation? i.e. How is the conversion > used? > > I am fine with the rest of the analytical VaR > material. It is really just that one little > section on conversion to a standard normal > distribution. The book probably talks about converting any normal distribution to standard normal distribution and back. It might be helpful to come up with VaR values for any normal distribution using z-values of standard normal and the above conversion.
Yes, I understand why it might be helpful, but it sounds like it isn’t necessary to know it for the exam. Based on what CPK123 says above, I doubt we would ever be using the actual underlying data to calculate standard deviation (e.g. calculating standard deviation from monthly returns) rather than being provided standard deviation for different assets and then having to calculate the portfolio standard deviation, and using that rather than having to convert the raw data.