RMS / statistics question

[shluffer]

in RMS an assumption is made that gross of per risk, ceded to per risk and net of per risk are all three beta distributions. Would anyone happen to know if the sum of two beta distributions is a beta distribution? If not, what assumption is made in RMS to get this to work?

[diegol]

> Would anyone happen to know if the sum of two beta distributions is a beta distribution?

I view of this: http://www.groupsrv.com/science/about10341.html I would guess not.

[BassFreq]

I don't believe that If X and Y are beta distributed then their sum is also beta distributed. However, considering the flexibility of the Beta, I wouldn't be surprised if the sum is close enough to beta distributed.

[shluffer]

Thanks for the link. It answers the question to my satisfaction.

[Ganymede]

Here is an easy way to see it. Beta distributions go between 0 and 1. The sum of two betas go between 0 and 2.

[BassFreq]

Moran!