Define the hazard rate, then apply it to find the unconditional probability that a credit asset defaults.
The idea
The hazard rate, λ, is the instantaneous default intensity. The quantity λ × dt is the probability of default in the next short interval dt, conditional on survival to that moment. With a constant hazard rate, the survival probability to time t is S = e^(−λt). The cumulative default probability is then PD = 1 − e^(−λt). Between times t1 and t2, the unconditional probability of default is PD = e^(−λt1) − e^(−λt2). If the hazard rate changes by period, add λ × length across the periods in the exponent. A common market approximation is λ = spread ÷ (1 − recovery rate). The usual errors are using λt as the default probability, which overstates it, and confusing the cumulative probability to a date with the probability in one particular year.
Formulas
Survival probability with a constant hazard rate
S(t) = e^(−λt)
- λ
- the constant hazard rate per year
- t
- the time in years
Cumulative PD to t is 1 − e^(−λt). Using λt as the default probability overstates it; with changing hazard rates, add λ × length across periods in the exponent.
Unconditional default probability between two dates
PD(t₁ to t₂) = e^(−λt₁) − e^(−λt₂)
- λ
- the constant hazard rate
- t₁, t₂
- the start and end of the period, in years
The fall in survival over the period. Do not confuse it with the cumulative probability of default to t₂.
Hazard rate implied by a credit spread
λ = spread ÷ (1 − R)
- spread
- the CDS spread or bond yield spread over the risk-free rate
- R
- the expected recovery rate
An approximation that treats the whole spread as compensation for expected default loss. The result is a risk-neutral rate, higher than real-world default frequencies.
A practice question
Which statement defines the hazard rate, lambda, of a credit asset?
- Lambda is the probability today that the asset defaults at any time before maturity.Why not: Describes the cumulative default probability, which is 1 - exp(-lambda x t) under a constant hazard.
- Lambda x dt is the chance of default in the next short interval dt, given survival so far. - correct
- Lambda is the fraction of the exposure lost when default occurs.Why not: Describes loss given default, which is 1 minus the recovery rate.
- Lambda is the annual probability that the rating will fall by at least one notch.Why not: Describes a downgrade probability from a transition matrix, not an instantaneous default intensity.
The hazard rate is the conditional default intensity: given survival to t, the probability of default in the next dt is about lambda x dt.
A harder one
A firm has a constant hazard rate of 2% per year. The probability that it defaults at some point within five years is closest to:
- 10.00%Why not: Multiplies lambda by time (2% x 5), which ignores that default probabilities cannot be added over time when survival matters.
- 90.48%Why not: Reports the five-year survival probability, exp(-0.10), instead of the default probability.
- 9.52% - correct
- 1.76%Why not: Reports only the unconditional probability of default in year 5, exp(-0.08) - exp(-0.10).
Cumulative default = 1 - exp(-0.02 x 5) = 1 - exp(-0.10) = 1 - 0.9048 = 9.52%.
The hardest tier
An issuer has a constant hazard rate. Its probability of default within five years is 20%. The unconditional probability that it defaults during year 4 (between t = 3 and t = 4) is closest to:
- 4.5%Why not: Stops after backing out lambda = -ln(0.8)/5 = 4.46% and reports it as the year-4 probability.
- 16.3%Why not: Reports the cumulative default probability through year 4, 1 - 0.8^(4/5), not the single year.
- 3.8% - correct
- 4.0%Why not: Divides the 20% cumulative probability by five, assuming evenly spread unconditional probabilities.
Survival to year 5 is 0.80, so lambda = -ln(0.80)/5 = 0.04463. Year-4 unconditional PD = exp(-3 lambda) - exp(-4 lambda) = 0.8746 - 0.8365 = 3.8%.
In the app: 6 practice questions on this outcome, spaced review that brings it back before you forget it.
Start free →More in External and Internal Credit Ratings
- Describe the scales used by external rating agencies, how ratings are assigned, and how a rating connects to the chance of default.
- Distinguish default probabilities that are conditional from those that are unconditional, defining each.
- Define the recovery rate and, with it, compute the loan's expected loss.
- Contrast ratings built on a point-in-time basis with those built through the cycle, and explain each.
- Outline the options that can be used in place of the credit ratings issued by rating agencies.
- Weigh internal ratings against external ones, noting the strengths of each.
- Describe a rating transition matrix, calculate with it, interpret it, and explain what it is used for.
- Outline how credit default swap spreads, bond prices and stock prices tend to move when a credit rating is upgraded or downgraded.
- Explain where credit ratings have failed in the past and what difficulties arise when they are relied on for investment decisions.