TABLE OF CONTENTS Whereas PD has been the crucial issue discussed so far, we now turn to loss given default (LGD) and exposure at default (EAD). In recent years LGD has attracted a growing interest in research, due to the increasing perception of its importance in credit risk management. The key issues to be discussed for LGD are (1) how to define it, (2) how to measure it, (3) which factors affect it, and (4) whether it is correlated with PD. First of all, LGD can be defined in percentage terms as the fraction of the loan that is not recovered in the event of a default. Therefore it is equal to 1 minus the recovery rate (RR). Unfortunately, this implies that the definition and measurement of LGD depend on the definition of default. This issue has already been covered in the discussion of PD quantification, but it is equally relevant for LGD. Including past-due cases in the definition of default implies a higher PD but a lower LGD, since some obligors who delay payments for more than 90 days may sometimes be relatively sound borrowers that are later able to repay in full (i.e., LGD may be 0%), so it may be wise to model past-due LGD separately from LGD for other defaults (Schuermann 2005).
To avoid misunderstandings, we here define LGD as simply the ratio of the loss incurred on an exposure, relative to the exposure s EAD. This idea is important, since in the U.S.
