FRM: Beta distribution for loss given default (LGD)

[@bionicturtle]

@wooddog007 I can't recall, i may have manually input the formula/function or approximated with the cumulative CDF (by subtracting, you can get near). But, no matter: in EXCEL 2010, betadist() has been upgraded to beta.dist() which adds a paramater (true = CDF, false = pdf), so you might just upgrade to 2010? thanks, David

[@wooddog007]

great ... yes that is easier than putting in the density function manually !! I still use Excel 2007 at home and 2003 at work !!! Yikes. will look for that in the 2010 version. Thanks for your help.

[@wooddog007]

Dave, how do you draw the beta probability density function? Excel has BETADIST, but this only draws a cumulative probability distribution ... was wondering how you managed to construct these? Thanks.

[@PhiladelArt]

Is there a more or less easy way to arrive at alpha and beta values given a mean and st.dev of recovery rate? Many thanx!

[@ianjamesbarnett]

if you let your beta distribution have parameters less than or equal to 1 for both of them, then your distribution will be bimodal with peaks at 0 and at 1, which sounds like it might be a more realistic model. Sorry, I do not understand this economic context as I my field is in biostatistics, but it sounds to me like the choice of a Beta distribution is somewhat arbitrary, so a looser yet more complicated model (like a mixture of two betas) could also solve your problems.