Dear Allstat,
I'm just wondering if there is someone out there with experience of
modelling under Basel II who could shed some light.
Under the long run PD modelling: BIPRU 4.6.24 says, A firm must estimate
PDs by grade from long-run averages of 1 year default rates.
It also says in Bipru 4.4.65R that the PD of obligors in default must be
100%.
Does it therefore mean that when you are building a model and you are
looking at the long-run averages of 1 year default rates you do not
include
In your estimate/sample those accounts that are in default at the start
of the 1 year period.
By the same token, when you validate out of time looking at a 12 month
window, you do not include those accounts in default at the start of
that period?
Thanks for your help on this.
Best,
Jabu
**********************************************************************************
The information contained in this e-mail (and any attachment) is private and intended for the use of the addressee(s) only. Any unauthorised use or disclosure of information contained within it (or within any attachment) is strictly prohibited and may be illegal. Please let the sender know immediately if it has been sent to or received by you in error. Any opinions expressed in this e-mail are those of the individual and not Norwich and Peterborough Building Society or any of its subsidiaries. All Norwich and Peterborough Building Society e-mails are subject to monitoring.
N&P Mail Administrator
Norwich and Peterborough Building Society
Peterborough Business Park, Lynch Wood, Peterborough, PE2 6WZ.
Principal Office: Peterborough Business Park, Lynch Wood, Peterborough, PE2 6WZ.
Norwich and Peterborough Building Society is authorised and regulated by the Financial Services Authority (registration no. 150965) and entered in the Financial Services Authority's register. www.fsa.gov.uk/register
Participant in the Financial Ombudsman Service scheme.
The Society provides financial advice on regulated and non-regulated products.
http://www.nandp.co.uk
You may leave the list at any time by sending the command
SIGNOFF allstat
to [log in to unmask], leaving the subject line blank.
|