Credit Risk: Default Probability, LGD, and Credit Models
Probability of default, loss given default, exposure at default, credit ratings, structural models (Merton), and reduced-form models.
Study depth
6
structured sections
Exam lens
20%
topic blueprint weight
Access
Premium
visual summary included
Concept visual
Expected Loss Building Blocks
Credit risk usually decomposes into probability of default, loss given default, and exposure at default.
Why it matters
When the exam changes one component, trace the directional effect on expected and unexpected loss separately.
Credit Risk: Default Probability, LGD, and Credit Models
Credit Risk Components
Credit risk loss depends on three variables:
Expected Loss = PD ร LGD ร EAD
Where:
- PD (Probability of Default) โ the likelihood a borrower defaults within a given time horizon
- LGD (Loss Given Default) โ the percentage of exposure lost if default occurs (1 โ Recovery Rate)
- EAD (Exposure at Default) โ the total amount exposed when default occurs
Expected vs. Unexpected Loss
- Expected Loss is priced into spreads and provisioned for โ it is a cost of doing business
- Unexpected Loss is the volatility around expected loss โ this is what requires capital
Capital = f(Unexpected Loss) = f(ฯ of portfolio credit losses)
Probability of Default
From Market Data โ Credit Spreads
Credit spread = yield on risky bond โ risk-free yield
Approximate PD (annual): PD โ Spread / (1 โ Recovery Rate)
Example: If a bond has a 200 bp spread and expected recovery is 40%: PD โ 0.02 / 0.60 = 3.33%
This is the risk-neutral PD, which includes a risk premium. The physical (actual) PD is typically lower.
From Historical Data โ Rating Agencies
Rating agenc
...
Premium Library Access Required
The first two study notes stay free. The remaining 11 notes are reserved for premium members, with full reading access, charts, and deeper exam guidance.
Unlock Premium Notes