Impact LGD from period 1 to period 2

[KlarinettSax]

Hello,

I have a question. The IRB approach to calculate RWA is RWA=12,5*K*EAD. The formula to calculate K is K = a * LGD * f(PD,conf.interval) * f(PD,maturity).

I want to calculate the effect on RWA from changes in EAD, LGD and PD. Let's say I have period 1 and period 2. My plan is to first calculate total RWA in period 1 and then total RWA in period 2. RWA period 2 minus RWA period 1 is the total RWA change. I now want to determine how much of this change in RWA comes from changes in EAD, LGD and PD separately. For example, if I want to determine how much of the total RWA change comes from a change in LGD, can I then just keep all values from period 1 in the calculation, and just have the LGD from period 2, in order to see how much the change in LGD has impacted RWA?

Many thanks!

 

[David Harper CFA FRM]

Hi @KlarinettSax Here is the IRB sheet from our Basel workbook https://www.dropbox.com/s/12goncaxipw9ebo/1128-basel-irb.xlsx?dl=0
The exposure and LGD are linear, so they are easy (or easier). The challenge is that PD enters the Merton/ASRF function in a very non-trivial way (formula is in the XLS). It is interesting to me that you want to solve for, essentially, marginal RWA (ΔRWA/ΔPD), but I do not recall reading specific examples of this. I'd probably go the simulation route. I hope the XLS is helpful, good luck!