What's new

Stress testing and other risk management tools

lowhueyyi

New Member
Subscriber
Thread starter #1
Hi David, in your note (chapter 2) for (Siddique and Hasan stress testing pg 13), the value for two year cumulative rate is 7.44%. However, the GARP official text (pg 303) gave a value of 10.61%. How do we get the value for two year cumulative loss rate and which is the correct value?
 

David Harper CFA FRM

David Harper CFA FRM
Staff member
Subscriber
#2
Hi @lowhueyyi We did not run the calculations on that exhibit but rather just used (copied) the source text, as below:


... but the text edition must have been updated (it would not be like GARP to run calculations, not even ) because quickly I can see that 10.x% does appears to be correct. See below. All i did was calculate the 2-year cum pd = 1 - (1-PD1)*(1-PD2), then final column is the stressed loss = Bal * 2-yr cum PD * LGD. Then 209.81/2000 = 10.49%; I did this in ~3 min so i could have a typo .,..


Can I trouble you for a screen shot? I think the source must have been updated etc

cc: @Nicole Seaman

Thanks,
 

lowhueyyi

New Member
Subscriber
Thread starter #3
Hi @David Harper CFA FRM , below are the images from the 2019 edition GARP text.
1565398972480.png 1565398995432.png

Btw do we have to know how to calculate cumulative loss rate in part 1? I am not sure where this formula is covered in part 1.
 
Last edited:

David Harper CFA FRM

David Harper CFA FRM
Staff member
Subscriber
#4
Hi @lowhueyyi Thank you, I did not realize it had been thusly updated. Re: "do we have to know how to calculate cumulative loss rate in part 1? I am not sure where this formula is covered in part 1" ... this is typical of the FRM. This exact formula is not exactly covered in Part 1; it is certainly not assigned in this reading which has no quantitative action verbs (calculate, compute). However, it's almost fair game in my opinion in the total context of Part 1. Omitting the "stressed," notice how the calculation for each balance is:
  • Loan * cumulative 2-year PD * LGD
  • which is then summed over the buckets; i.e., as a portfolio
That's just expected loss (EL ) = Exposure * PD * LGD, which is covered; I guess you could argue cumulative PD is truly in P2 ...

Now that you identified this, I think we (our note) is wrong to omit the calculation, is to suggest I think a P1 should know this, even if it's strictly borderline, because that's how GARP runs the FRM, their borders on the questions are fuzzy ... I hope that's helpful

THANK YOU for alerting me to this discrepancy!
 
Top