Gregory CH:7 Exposure definitions EE and PFE

[pascalb]

Hello, although feeling comfortable with the topic, there is this nagging uncertainty about EE and EPE.
(1) Is EE an exposure number that the fin inst. has for EACH cp (ie the average of all exposure values with a particular cp) and is EPE an exposuure that a fin inst. has across ALLcps (average of all exposures across all time horizons).
(2) Is EE at a point in time?

 

[Headfield]

@pascalb It is generally quite well described in the study notes.

In a nutshell:

EE = expected exposure = expected positive credit exposure (credit exposure is always positive or zero, as negative values have no counterparty risk i.e. no credit exposure -> if no one has to pay you money, no one can fail in doing so) = the expected exposure your counterparty(ies) owe you at a specific point in time

That means you can build up a curve of EEs over points of time in the future

EPE = the weigthed average of EE; if your points in time in the future have the same distance, then this is a simple average of all your EEs

PFE (at least in your threadname) = potential future exposure, that has a comparable concept as EE, i.e. specific values for specific points of time in the future, however, it is a worst case (with a confidence level, somehow comparable to VaR) value

EE(t = 1) --> the expected value of EE at point of time = 1, like the mean of all the scenarios you would think possible for this point of time
PFE(t=1) --> in case the market develops far better in your favor than you expected for point of time = 1, you might end up having your counterparty(ies) owing you much more than the EE; this is of course nice from your potential revenue perspective, but at the same time strongly increases the potential loss (or rather not revenue) in case your counterparty(ies) defaults; therefore "worst case"