Hong Kong Baptist University · FACULTY OF MATHEMATICS

MATH7910 Chap.8 Default Probability and Credit Risk

- one subject, every graph, every model, every mark
6 Chapters3-page Bible
Our own words - no uploaded lecturer files
Updated for this semester
Chapter 8 of 9 · MATH7910

Default Probability and Credit Risk

Conditioning changes the denominator

Cumulative default probability measures default by a horizon. Interval probability subtracts cumulative values at two horizons. Hazard conditions the interval default on survival to its start, so the denominator excludes earlier defaults.

Survival and cumulative default sum to one.

Probability becomes loss through recovery

Recovery is the fraction retained after default; loss given default is its complement. Expected loss combines exposure, default probability and loss given default on aligned horizons. Historical frequencies and risk-neutral probabilities inferred from prices answer different questions and need not match.

In this chapter

What this chapter covers

  • 01

    Credit ratings

  • 02

    Cumulative default probability

  • 03

    Survival probability

  • 04

    Conditional hazard

  • 05

    Recovery and loss given default

  • 06

    Historical and risk-neutral estimates

Worked example · free

Calculate a conditional hazard

Q [6 marks]. The mark allocation here structures AskSia practice and is not a university assessment scheme. Q one is 3% and Q two is 7%. Find year-two hazard conditional on survival.
As cumulative default probability increases across horizons, the matching survival probability decreases so the two always sum to one.As cumulative default probability increases across horizons, the matching survival probability decreases so the two always sum to one.Survival fallsDefault rises
As cumulative default probability increases across horizons, the matching survival probability decreases so the two always sum to one.
  • 2Find interval default.
  • 2Find survival denominator.
  • 2Condition and interpret.
Interval default is 4%. Survival to year one is 97%. Hazard is 4%/97%, approximately 4.124%.
Sia tip — Use survivors, not the original population, in the hazard denominator.
Glossary

Key terms

Survival Probability
The probability that default has not occurred by a specified horizon.
Cumulative Default
The probability that default has occurred by a specified horizon.
Hazard Rate
Default probability over an interval conditional on survival to its start.
Loss Given Default
The fraction of exposure lost after accounting for recovery.
FAQ

Default Probability and Credit Risk FAQ

Why is hazard larger than interval default probability?

The numerator counts the same new defaults, but the hazard denominator excludes entities that defaulted earlier. When earlier default is small, the difference is small; when it is large, conditioning matters more.

Can cumulative default probability decrease?

No, not in an absorbing default model. Extending the horizon can only add possible default events. A decreasing cumulative series signals a transcription, denominator or category error.

Why can historical and market-implied probabilities differ?

Historical estimates describe observed frequencies. Market prices also reflect risk premia, liquidity and recovery assumptions. A risk-neutral probability is designed for pricing, so it need not equal the best forecast of actual default frequency.

Study strategy

Exam move

Use a population diagram for every default-probability question. Begin with the original cohort, mark cumulative defaults by each horizon, and write survivors as one minus cumulative default. Obtain default during an interval by subtracting endpoint cumulative probabilities. For a hazard, divide those new defaults by the survivors available at the interval start.

Write the conditioning event in words before the fraction and keep numerator and denominator in the same percentage or decimal scale. Test the resulting series: cumulative default cannot decrease, survival cannot increase in the absorbing model, and every conditional probability must remain between zero and one. Next, attach recovery.

Write recovery rate and loss given default as complements, then multiply exposure, probability of default and loss given default only when their horizons and exposure definitions align. Distinguish expected loss from worst-case loss and from a tail credit-risk measure. Compare a historical rating-table estimate with a market-implied probability and explain why risk premia, liquidity and recovery assumptions can separate them.

Do not describe the pricing probability as the best forecast without evidence. Practise one bond-spread approximation and list the assumptions needed before dividing spread by one minus recovery. Then work one structural example conceptually: equity resembles an option on firm assets, and default occurs when asset value cannot cover debt at the horizon. Keep that model separate from observed cohort frequencies.

During revision, classify each requested quantity as cumulative, interval, conditional, survival, recovery or expected loss before calculating. If the answer is wrong, inspect category and denominator before arithmetic. Finish by reconstructing the relationships from a blank sheet, including Q plus V equals one and loss given default equals one minus recovery.

Create a small synthetic rating table and check every row for monotone cumulative default before using it. Convert the table into interval probabilities, then into conditional hazards, keeping the original cohort size visible. Change recovery while holding default probability fixed and explain why expected loss moves even though credit-event frequency does not.

End with a short comparison of historical, spread-implied and structural estimates, naming the evidence and assumptions behind each rather than selecting one as universally correct. End with a unit-cohort exercise. Imagine ten thousand bonds and convert each probability into a count of defaults or survivors. Perform the interval and hazard calculations with counts, then translate back into percentages.

This makes the conditional denominator concrete and exposes impossible results immediately. Apply recovery to the resulting defaulted exposure only after the probability stage is reconciled, keeping frequency and severity as separate sources of loss. Translate one conditional hazard back into an unconditional interval probability by multiplying by survival.

The reversal confirms that numerator and denominator describe the same cohort.

Working through Default Probability and Credit Risk in MATH7910? Sia is AskSia’s AI Mathematics tutor — ask any MATH7910 Default Probability and Credit Risk question and get a clear, step-by-step explanation grounded in how MATH7910 is taught and assessed. Read this chapter free, then take your hardest questions to Sia.

A+Everything unlocked
Unlocks this Bible + all 16 of your subjects at Hong Kong Baptist University, plus 1,000+ more Bibles across the AskSia library.
Sia - your MATH7910 tutor, unlimited, worked the way the exam marks it
The full 3-page Bible + practice bank with worked solutions
Chrome extension - sync your LMS so Sia knows your deadlines
Bilingual EN / Chinese on every Bible and every Sia answer
$0.99 Trial
30-day money-back · cancel in one tap · how it works