Under GIPS, would the difference between the 2 highest and 2 lowest returns in a composite be an acceptable method for calculating internal dispersion?
Wasn’t there an answer for std deviation?
i thought the two others were variability in the composite’s return itself… top two vs. bottom two threw me, but i went with it.
cdogstu77 Wrote: ------------------------------------------------------- > Wasn’t there an answer for std deviation? That is what I tought
westbruin Wrote: ------------------------------------------------------- > i thought the two others were variability in the > composite’s return itself… > > top two vs. bottom two threw me, but i went with > it. That was my thought as well.
but i thought it said std dev of diff between composite and benchmark . i think 2 of the choices said composite and benchmark -evidently not internal dispersion related.so i chose the third one.
Dsylexic Wrote: ------------------------------------------------------- > but i thought it said std dev of diff between > composite and benchmark . i think 2 of the choices > said composite and benchmark -evidently not > internal dispersion related.so i chose the third > one. this.
The answer for standard deviation was time-series standard deviation of total composite returns, which is NOT acceptable. The correct answer was the range (diff between 2 highest and 2 lowest returns here).
the standard deviation answer was dispersion of portfolio’s with benchmark, not dispersion of individual portfolio returns. Correct answer was difference between avgs of 2 highest/lowest, which is essentially the range
Yeah, 2 of them were composite dispersions so I went with the only internal dispersion measure.
remember dispersion is related to returns of portfolios within the composite not the returns of portfolios in the composite relating to a benchmark.
The standard deviation was the sigma of the composite time series and not the dispersion of the composite components. I am very sure about that one. I went with the average of the two as well but I had not heard of this method. I was just familiar with high/low (which i guess is what this was), Interquartile and sigma.
Dwight Wrote: ------------------------------------------------------- > westbruin Wrote: > -------------------------------------------------- > ----- > > i thought the two others were variability in > the > > composite’s return itself… > > > > top two vs. bottom two threw me, but i went > with > > it. > > > That was my thought as well. i looked at it so much too… hope we’re not wrong LOL
Not sure if the method is relevant as long as it’s an internal measure and it’s disclosed.
i knew that this question looked too easy
Slash Wrote: ------------------------------------------------------- > Not sure if the method is relevant as long as it’s > an internal measure and it’s disclosed. Per Schweser: “The GIPS handbook identifies the following acceptable methods for calculating internal dispersion: -rang of annual returns -high and low annual returns -interquartile range -standard deviation of equal-weighted annual returns -asset weighted standard deviation of annual returns” I guess you are technically right though that they don’t specifically say that you are NOT ALLOWED to use a different method of internal dispersion so that might be OK.
I guess that is why GIPS indicates that a description of the methodology for calculation is available upon request. Sounds like we are correct with the high/low.
some of those early ones seem suitably vague, now that i see them… but i didn’t think so going in (i.e. no other good answer)
The average between 2 high and 2 low seemed alot like the so called “inter-quartile range” they talked about
average two high two low, we care about the dispersion of portfolio returns not the composite returns