delaurentis - chapter 2 - recovery

[farahm]

Can someone elaborate on the following text below: I understand how recovery would be generated at couterparty's position but not sure I'm comfortable with the term globally? so are we looking at the average of a conglomeration of positions to get a pre-transaction value for recovery or 1-LDG?

1. Recoveries are mostly calculated globally at the counterparty’s position such that their reference to the original contracts, collaterals, and guarantees is mostly lost. Then, ‘top down’ procedures are normally used to outline the average loss given default rates for a related set of facilities and guarantees.
   
   
   Maybe I'm thinking too deep on this one ... thanks
   

 

[Daniel26]

I have read it many times now and still not 100 percent sure. But I guess the main point he makes is that Recovery depends heavily on the jurisdiction and type of contract and that this information is lost if you look at a Recovery from a generalized or global point of view. So from a model perspective (regarding estimation of Recovery) it would be very hard to incorporate the specific characteristics of each jurisdiction for example to arrive at "one real Recovery".

Happy to discuss.

 

[bpdulog]

Can someone elaborate on the following text below: I understand how recovery would be generated at couterparty's position but not sure I'm comfortable with the term globally? so are we looking at the average of a conglomeration of positions to get a pre-transaction value for recovery or 1-LDG?

1. Recoveries are mostly calculated globally at the counterparty’s position such that their reference to the original contracts, collaterals, and guarantees is mostly lost. Then, ‘top down’ procedures are normally used to outline the average loss given default rates for a related set of facilities and guarantees.
   
   
   Maybe I'm thinking too deep on this one ... thanks

I think the key takeaway here is that, assuming this is a large bank with millions of transactions, you won't be able to figure out a specific positin's recovery rate because that data gets lost. The reason why, I think, is that if a bank has 1000s of exposures with a single client, then they will just take the average LGD for that client as opposed to calculating each individually