Reply To: Can EPM’s be used to override the deals PnL Detail ?

Israr Ahmed



 EPM returns the MtM ( at a possible 3 granularities – see above )

 Endur then maps this onto the PnL Detail ( which has more granularity so not sure how it does the mapping ) as the “Total PnL” value .

I think Endur also then looks at the Realised setting ( Payment v Delivery ) and then must have an Instrument Type specific logic that then generates a Realised PnL value based on this setting which is subtracted from the “Total PnL” to get both “Realised PnL” and the remainer = “Unrealised PnL”

Results Calculator

 From what I can understand , this uses the EPM as input ( or can calculate value itself ) and then can write directly into the PnL Detail ( and other results ) without the need for Endur to do the mapping from MtM to PnL Detail.

 This also allows us to :
o Change the structure of PnL Detail ( ie for Y/Q Futures add monthly fake monthly profiles )

o Write into Realised and Unrealised directly ( rather than relying on Endur to do the Realised split)


 Is the outline above correct ?

 Does the Results Calculator need an EPM as input ?

 The EPM can also be used to send back Volumes and Greeks … is this only if the Results Calculator is used ?

 Usually EPM model is required when the value is not ontainable or calculatable in Endur itself. However there are lots of examples where we have all the numbers we need in Endur , its simply that the PnL Detail is not correct ( Futures is good example) . In that case it seems we have two options to correct the PnL Detail :

o Results Calculator above
o Create a UDSR that calls PnL Detail and change the PnL Detail in memory within the UDSR script ( which does nothing apart this ) . Have seen this at both Citi and AXPO . It allows you to change the PnL Detail ( as long as you have the numbers you need to calculate the value in Endur ) without needing any OpenComponents licences. Does anyone see anything fundamentally wrong with this approach ?

Download PDF version

This field is for validation purposes and should be left unchanged.