In template C 103.00 ”Details on exposures in High Default Portfolio”, column 230 and 240, it is demanded the RWA* and RWA**. For those purposes, we need to calculate a PD* and PD**. Our question is about how PD* must be estimated. In this sense we have read the specifications and we need more clarifications or be sure our understanding is correct.

We have two ways:

1. For the bucket Portfolio ID - Rating (which belongs to the internal master scale) there are a Default Rate and a regulatory PD for 2015 which is calculated as an average. Therefore, for each bucket it can have different PDs which come from different models and levels. In this first option, we consider that we have to take every PD of each model and level and verified if PD + N-1(q)*(PD*(1-PD)/n)^0.5 is greater than Default Rate, being n the cases considered in the PD calibration process for each model/level. After that, take the minimum of the PD which satisfied the condition and this one will be the PD* to calculate RWA* whenever PD* is greater than official regulatory PD. Is this way correct for the calculus of PD*?
2. The other option is a top down approach. We have to use the average PD (PDa) for the rating grade (including all models in the bucket), verified if PDa + N-1(q)*(PDa*(1-PDa)/n)^0.5 is greater than Default Rate. If yes and PDa is greater than the average regulatory PD also, PD*=PDa, if not, PD* = regulatory PD and RWA = RWA*.

The instructions for c250 to c280 of template C 103.00 in Annex IV of the Draft ITS on Supervisory Reporting for Institutions for benchmarking the internal approaches (ITS on benchmarking) require the hypothetical PD* to be determined at rating grade level. Consequently the PD assigned to each specific obligor grade or pool shall be considered for supervisory reporting purposes  shall be considered, regardless of the underlying rating system. In particular, the average PD and average default rate shall be used as proposed under option 2 (n referring to the number of obligors in the benchmark portfolio one year before the reference date). In general, the same rating scale as reported in the Template C 08.02 of Annex I of Regulation (EU) No 680/2014 - ITS on Supervisory Reporting of institutions (ITS on reporting) shall be used.      

DISCLAIMER:   The present Q&A on Supervisory reporting is provisional. It will be reviewed after the Implementing Regulation is in force and published in the Official Journal. The text of the Implementing Regulation may differ from the text of the draft ITS to which this Q&A refers.

Update 26.03.2021: This Q&A has been updated in the light of the most recent amendments to the ITS 2016/2070 on Supervisory Benchmarking.