FRM Exam Part II · Estimating Default Probabilities

Recovery Rates and Their Link to Default

Q: How is the recovery rate measured for bonds?

It is usually the bond's market price about 30 days after default, as a percentage of face value. Market prices are used because final bankruptcy payouts take a long time to settle.

Q: What is the difference between loss given default and recovery rate?

They add to 100% of exposure. LGD = 1 − recovery rate. Recovery is what you get back, LGD is what you lose, and LGD is the input used in expected loss.

Q: Why are recovery rates and default rates negatively correlated?

In downturns many firms default together and asset values and demand are weak, so creditors recover less. In good times defaults are fewer and recoveries tend to be higher. This is an empirical tendency, not a fixed law.

Q: How does seniority affect recovery?

Senior secured claims are paid first and have collateral, so they recover most on average. Senior unsecured, subordinated and junior subordinated debt follow in that order, with lower average recoveries.

Updated 11 October 2026 · Fact-checked

The recovery rate is the share of a bond's face value that creditors get back after default, usually measured by the market price about 30 days after default. LGD = 1 − recovery rate. Senior secured debt recovers more than junior debt, and recovery rates tend to fall when default rates rise.

Understand Recovery Rates and Their Link to Default

When a borrower defaults, creditors rarely lose everything. The part they get back is the recovery rate. It is quoted as a percentage of face (par) value. If a bond with face value $100 is worth $40 after default, the recovery rate is 40%.

For bonds, recovery is usually measured as the market price of the bond shortly after default (commonly about 30 days after), expressed as a percentage of face value. Market prices are used because the final cash flows from bankruptcy can take years. Note that recovery is a share of face value, not of the bond's price before default and not of the claim including accrued interest in every definition. Check the wording of the question.

Loss given default (LGD) is the mirror image: LGD = 1 − recovery rate. A 40% recovery rate means a 60% LGD. Expected loss combines default probability and LGD, so any move in recovery changes the loss even if default probability is unchanged.

Recovery depends on seniority and security. Senior secured debt is paid first and is backed by collateral, so it recovers most. Senior unsecured comes next, then subordinated and junior subordinated debt. Historical averages on this ladder fall steadily as you move down, though any single case can differ. Industry, the business cycle and the legal system also matter.

The key link is that recovery rates and default rates are negatively correlated. In recessions many firms default at once, and the assets sold by creditors are hit by weak demand and fire sales, so recoveries are low. In good years defaults are few and recoveries are higher. This matters for risk: assuming a fixed recovery understates losses in a downturn, because high PD and high LGD occur together. Risk models that treat recovery as independent of default will understate tail losses.

Key formulas to remember

Recovery rate (bonds): Recovery rate = Post-default market price ÷ Face value; Price is typically observed about 30 days after default. Expressed as a % of face value.
Loss given default: LGD = 1 − Recovery rate; Applies per unit of exposure. A 35% recovery gives a 65% LGD.
Expected loss: EL = PD × LGD × EAD; EAD is exposure at default. Use LGD, not recovery, in this formula.
Seniority ordering: Senior secured > Senior unsecured > Subordinated > Junior subordinated; Ordering of average recoveries and priority of claim. Individual cases can differ.
Default–recovery relationship: Corr(default rate, recovery rate) < 0; Negative on average. It is an empirical tendency, not an exact rule.

How to solve Recovery Rates and Their Link to Default questions

Use this method for any question on recovery rates, LGD or their link to default.

1Identify what is given: a recovery rate, an LGD, a post-default price, or a loss amount.
2Check the base: is it a percentage of face value, and is the price measured after default?
3Convert between recovery and LGD with LGD = 1 − recovery.
4If PD and exposure are given, compute EL = PD × LGD × EAD.
5For seniority questions, rank claims: senior secured, senior unsecured, subordinated, junior. Higher rank means higher expected recovery.
6For correlation questions, remember the sign is negative: more defaults go with lower recoveries.
7State the implication: ignoring this link understates losses in downturns.
8Check that your answer is in the right units and the option matches the question (recovery or loss).

Quickest way: Flip, multiply, then sanity-check

When to use it: For numerical MCQs with recovery, LGD and expected loss, and for conceptual options on correlation.

Write LGD = 1 − R at once. Many wrong options are the recovery rate itself.
Multiply PD × LGD × EAD in that order.
For concept options, eliminate any answer saying recovery and default rates are positively correlated or independent.
Eliminate any option ranking junior debt above senior debt.
Confirm the answer asks for a loss, not a recovery.

Common mistakes in Recovery Rates and Their Link to Default

Using the recovery rate in the expected loss formula instead of LGD.
Both numbers are given and look similar.
Fix: Always convert first: LGD = 1 − R, then EL = PD × LGD × EAD.
Saying recovery and default rates are positively correlated.
Students link 'default' with 'recovery' as both rising with credit events.
Fix: Remember the stress picture: many defaults, depressed asset values, low recovery. The correlation is negative.
Treating the recovery rate as a percentage of the pre-default bond price.
Price is the number students usually work with.
Fix: Recovery is quoted as a percentage of face value, with the price taken after default.
Assuming seniority guarantees a higher recovery in every individual case.
Averages are read as rules.
Fix: Say that seniority raises average recovery and priority of claim. Single defaults can deviate.
Treating recovery as a fixed constant in risk models.
Textbook examples often assume a flat 40%.
Fix: Know that fixed recovery ignores the negative link with default and understates tail loss in downturns.

Worked examples

Example 1

A senior unsecured bond with face value $100 default trades at $38 about 30 days after default. A bank holds $20 million of this bond. The bank's one-year PD estimate for such issuers is 2%. Calculate the recovery rate, LGD and expected loss.

Show the solution

Recovery rate = 38 ÷ 100 = 38%.
LGD = 1 − 0.38 = 62%.
EAD = $20 million.
EL = 0.02 × 0.62 × 20,000,000.
0.02 × 0.62 = 0.0124.
0.0124 × 20,000,000 = $248,000.

Answer: Recovery rate 38%, LGD 62%, expected loss $248,000.

Example 2

In a recession, a risk manager's model uses a fixed 40% recovery rate for all defaults. Historical data show that when the default rate doubles, average recovery falls to 30%. For a $10 million exposure with PD of 4% (the doubled rate), how much does expected loss rise if the model uses 30% recovery rather than 40%?

Show the solution

With 40% recovery: LGD = 60%. EL = 0.04 × 0.60 × 10,000,000 = $240,000.
With 30% recovery: LGD = 70%. EL = 0.04 × 0.70 × 10,000,000 = $280,000.
Increase = 280,000 − 240,000 = $40,000.
Percentage rise = 40,000 ÷ 240,000 = 16.7%.
Interpretation: the fixed-recovery model understates loss because high PD and low recovery occur together.

Answer: Expected loss is $40,000 higher (about 16.7%) when recovery falls to 30%; the fixed 40% assumption understates loss in a downturn.

Exam tips

Convert recovery to LGD before anything else; wrong options often use the unconverted number.
For correlation questions, the safe answer is negative, and the reason is shared stress in the credit cycle.
Know the seniority ladder in order and be ready to apply it to a short case.
Read whether the question asks for a recovery or a loss and give that figure.
In case questions, link a fixed-recovery assumption to understated tail risk.

Practice questions from Estimating Default Probabilities