Let's say I have a risk factor that is defined between [0,1], such as recovery rates. Assuming I have daily data, I can estimate the "daily VaR", i.e. the tails over 1 day period, since the data is available every day. I would like to estimate the changes over a X-day period. Therefore, I am trying to see whether I can apply some sort of scaling (i.e. a version of square root of time) to the 1-day estimates? Is there a straightforward way to do it for Beta distributions? Thanks!