Conditional and Unconditional Default Probabilities in Lifetime LGD

Summary

The document raises a modeling question about which form of probability of default belongs in a term structure used to calculate lifetime loss given default. It distinguishes, at a high level, conditional default probabilities from unconditional probabilities and notes that their values may evolve differently across time horizons. The author is unsure how to reconcile those definitions and wonders whether a declining conditional probability reflects survival as a customer continues to pay.

The text contains no worked calculation, formal definitions, or resolution, so it does not specify a preferred approach to LGD estimation. Its useful contribution is to highlight that probability conventions matter when constructing a term structure and interpreting changes over time. Readers would need a credit-risk framework that defines each period's conditioning event and explains how default probabilities are aggregated before applying them in a lifetime loss calculation.

Key ideas

• Lifetime LGD estimation may use a term structure of default probabilities.
• Conditional and unconditional default probabilities represent different probability conventions.
• The choice of convention affects how a term structure is interpreted over time.
• The document asks whether survival and good payment history explain a declining conditional probability, but gives no resolution.

Tags

# conditional or unconditional probability in the calculation of LGD


# conditional or unconditional probability in the calculation of LGD












lifetime LGD is calculated by the way of a term structure of PD. But this PD at the end are calculated using conditional or unconditional PD? I have found a Lot of different And confusing definition about that. If conditional PD are used the PD tend To decrease. While, I guess, it should decrease due that the Client has been a good payer.

Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.