Loss given default (LGD)

Cure, re-default, and recovery timing

Build transparent default episodes, discount workout cash flows, and monitor cure and repeat-default outcomes.

Illustrative episode-linkage policy and synthetic recovery cashflows. The example cure window is not universal; accounting ECL and prudential realized LGD follow different rules.

Purpose

Work through how a cure definition and re-default linkage change episode-level LGD, recovery timing, and monitoring. The 12-month probation below is synthetic and illustrates one declared policy; it is not a universal or recommended cure window.

Key terms

Hover over or focus a term for a quick definition, or open its glossary entry.

Key definitions

Cure means an account meets a stated return-to-performing rule and any required probation under the chosen methodology.
Re-default is a later qualifying default; whether it belongs to a prior or new episode depends on the applicable rule and model design.
Realized workout LGD compares exposure at a default date with discounted recoveries and relevant workout costs under a specified methodology.

Formula

PV(recoveries) = Σₜ Recoveryₜ / (1 + r)ᵗ
PV(workout costs) = Σₜ Costₜ / (1 + r)ᵗ
Illustrative realized LGD = (EAD₀ + PV(workout costs) − PV(recoveries)) / EAD₀
Redefault incidence by horizon H = eligible cured accounts re-defaulting within H ÷ cured accounts with full H follow-up

Worked example

Episode 1: EAD at Month 0 is $10,000. Discount recoveries of $3,000, $2,500, and $1,500 in years 1–3 and costs of $200 and $150 in years 1–2 at 5%. PV recoveries = $6,420.47; PV costs = $326.53; illustrative LGD = ($10,000 + $326.53 − $6,420.47)/$10,000 = 39.06%. A $1,600 recovery and $100 cost one year after a later Episode 2 default with EAD $4,000 give PV recovery $1,523.81, PV cost $95.24, and illustrative LGD 64.29%.

Synthetic event timeline under this article's declared example policy.
Month	Event	Episode treatment in this example
0	First default; EAD $10,000	Episode 1 begins.
6	Returns current	Illustrative 12-month probation begins, ending at Month 18 if status remains current.
15	Re-default before probation ends	Linked to Episode 1 by the declared example rule.
24	Returns current again	New illustrative probation runs through Month 36.
36	Still current	Cure completes in this synthetic example.
48	Later default; EAD $4,000	Episode 2 begins under this declared example rule.

Episode cashflows and present values, discounted at a synthetic 5% annual rate.
Episode	Default-date EAD	Cashflow timing	PV recovery	PV cost	Illustrative LGD
1	$10,000	Recoveries $3,000 (Y1), $2,500 (Y2), $1,500 (Y3); costs $200 (Y1), $150 (Y2)	$6,420.47	$326.53	39.06%
2	$4,000	Recovery $1,600 (Y1); cost $100 (Y1)	$1,523.81	$95.24	64.29%

Monitor cure, re-default and workout maturity together.
Metric	Example definition	Key interpretation
Cure rate	Episodes meeting cure criteria ÷ mature eligible episodes	Denominator and probation policy must be fixed.
Re-default incidence	Re-defaults by horizon H ÷ cured accounts with full H follow-up	Do not treat accounts without full follow-up as non-redefaults.
Time to cure	Elapsed time from default to return and completed probation	Report distribution/percentiles, not just the average.
Recovery curve	Cumulative recovery and cost by months since default	Compare vintages and note unresolved workouts.
LGD O/E	Observed loss divided by expected loss for matched mature episodes	Review counts, mix, censoring, and policy changes before interpreting movement.

WHY EPISODES NEED EXPLICIT RULES

A default episode is a governed grouping of delinquency, default, workout, cure, and potentially later re-default events used to measure losses. If event linkage is ambiguous, the same recovery can be counted twice, omitted, or attributed to the wrong exposure. State the unit of default, the cure trigger, any probation, what constitutes continuous performance, and whether re-default before probation completion remains linked to the original episode.

1. Fix the intended use

Identify whether the work supports IFRS 9 accounting, regulatory IRB LGD, stress testing, pricing, collections analytics, or internal monitoring. The word LGD does not make definitions or discounting interchangeable.

2. Declare an episode policy

In this teaching example, a return to current status begins a synthetic 12-month probation. A re-default before that period ends is kept in the current episode. A later default after a completed probation starts a new example episode. This is one transparent coding convention, not a universal rule.

3. Preserve an event timeline

Keep dated balances, arrears status, default flags, restructurings, cure dates, cash recoveries, collateral proceeds, and workout costs. Keep each raw event before applying episode transformations so a reviewer can reproduce the classification.

WORKED TIMELINE

Month 0 is the first default (Episode 1, EAD $10,000). The account returns to current at Month 6 and begins the illustrative probation through Month 18. A re-default at Month 15 happens before probation completion, so this example links it to Episode 1. The account returns to current again at Month 24; if it then stays current through Month 36, the example marks cure complete. A new default at Month 48 starts Episode 2 with EAD $4,000. Change the policy and this event path may be classified differently.

DISCOUNT CASH FLOWS TO THE DEFAULT DATE

For the simplified realized-LGD illustration, discount each recovery and qualifying cost back to its episode's default date. Do not net cashflows in nominal terms first and ignore timing. Episode 1 receives synthetic recoveries at year ends 1, 2, and 3, and incurs costs in years 1 and 2. The calculation gives PV(recoveries) $6,420.47 and PV(costs) $326.53. Subtract recoveries from EAD plus costs, then divide by default-date EAD: 39.06%. Episode 2 uses its own default date and EAD; its example result is 64.29%. Values are rounded to cents and percentages.

ACCOUNTING AND PRUDENTIAL METHODS ARE DIFFERENT

For IFRS 9, expected credit losses reflect probability-weighted contractual cash shortfalls, time value of money, and reasonable and supportable information under the standard. The applicable effective-interest-rate discounting and cashflow assumptions matter. The realized-LGD arithmetic above is not itself an IFRS 9 ECL calculation: it uses a single realized episode, a simplified flat discount rate, and fixed EAD.

For regulatory IRB LGD, use the applicable prudential definitions and supervisory guidance. EBA-GL-2017-16 includes specific treatment for realized LGD for cured exposures and subsequent defaults, with conditions in its paragraphs 135–136. That is a regulatory context and must not be generalized to every IFRS 9 estimate or internal analytics question. Basel framework data expectations also call for recovery timing and cost information.

Internal risk measures can define their own episode cohorts for forecasting or monitoring, but must label the convention, test alternative cure windows and episode rules, and keep it distinct from accounting and regulatory outputs.

MONITORING: DO NOT STOP AT ONE LGD

1. Cure rate

Eligible default episodes meeting the policy's cure criteria ÷ mature eligible episodes. Define whether the denominator is accounts, facilities, or episodes.

2. Re-default incidence

Re-defaults within a fixed horizon after cure ÷ cured accounts with full observation of that horizon. Exclude or explicitly model right-censored accounts that have not completed follow-up.

3. Time to cure

Track the distribution from default to return-to-performing and to completed probation; median alone can hide a long tail.

4. Recovery and cost curves

Compare cumulative discounted and undiscounted recoveries and costs at months-on-book from default, by vintage and segment.

5. LGD observed-to-expected

Compare realized and forecast loss on matched, mature episodes, while reporting sample counts, unresolved cases, mix, and uncertainty.

6. Episode quality

Monitor missing event dates, duplicate or negative cashflows, inconsistent EADs, re-default links, and episodes without a clear close or censoring status.

7. Segmentation and change

Review product, collateral, seniority, geography, collections path, and policy changes. A change in cure policy can shift measured LGD even if borrower risk has not changed.

IMPLEMENTATION STEPS

1. Build the dated event table

Keep one row per governed exposure event and retain source dates.

2. Apply the episode state machine

Store policy version, cure completion, and episode identifiers.

3. Link the cashflows

Attach each recovery and cost once to the correct episode, with source and cash date.

4. Discount with the chosen method

Match the rate and time basis to the measurement purpose.

5. Handle open workouts

Keep unresolved outcomes open or model censoring instead of assuming that no later recovery will occur.

6. Validate event assignments

Review sample accounts and compare results under plausible alternative cure/probation definitions.

7. Publish with context

Include purpose, population, maturity, source date, policy version, and metric owner.

REFERENCES AND SCOPE

The IFRS Foundation source describes IFRS 9's impairment requirements; the EBA guideline is an EU prudential IRB source; the Basel framework is a regulatory capital source. Consult the authoritative text that applies to the institution and jurisdiction. All amounts and dates in this article are synthetic, for education only.

Try a reproducible Python example

These examples use synthetic inputs and run locally in a browser worker. Edit the code, run it, and compare its output with the published example.

Runnable Python · local in your browser

Link re-default episodes and calculate discounted LGD

Apply a declared synthetic 12-month cure probation and compute two episode LGDs from bounded cashflow lists.

Editable Python example

from math import pow

# Synthetic example policy only: a re-default before 12 months links to the open episode.
cure_probation_months = 12
return_to_current_month = 6
redefault_month = 15
probation_complete_month = return_to_current_month + cure_probation_months
episode = 1 if redefault_month < probation_complete_month else 2
print("Month {} re-default linked to Episode {}".format(redefault_month, episode))

def pv(cashflows, annual_rate=0.05):
    return sum(amount / pow(1 + annual_rate, year) for amount, year in cashflows)

episodes = [
    {"name": "Episode 1", "ead": 10000, "recoveries": [(3000, 1), (2500, 2), (1500, 3)], "costs": [(200, 1), (150, 2)]},
    {"name": "Episode 2", "ead": 4000, "recoveries": [(1600, 1)], "costs": [(100, 1)]},
]
for item in episodes:
    pv_recoveries = pv(item["recoveries"])
    pv_costs = pv(item["costs"])
    lgd = (item["ead"] + pv_costs - pv_recoveries) / item["ead"]
    print("{}: PV recoveries={:,.2f} USD; PV costs={:,.2f} USD; illustrative LGD={:.2%}".format(item["name"], pv_recoveries, pv_costs, lgd))

Output

Run the code to see the output.

Runtime loads only after you press Run. Code runs in a separate browser worker with a 10-second limit and a 20,000-character output cap. Use synthetic examples only; browser execution is not a secure environment for secrets or sensitive records.

Assumptions

The example uses a 12-month cure probation purely for instruction; institutions and frameworks can define cure differently.
All synthetic cashflows are annual year-end amounts, discounted to each episode's default date at a flat illustrative 5%.
The example episodes use constant EAD within each episode and ignore collateral allocation, taxes, post-write-off interest, and other framework-specific adjustments.

Limitations

Cure windows, default definitions, and episode-linkage rules differ by purpose and jurisdiction; these examples are not prescriptive.
Unresolved workouts have incomplete future recoveries and costs; treating them as final outcomes can bias LGD.
Changing collection strategies, collateral mix, legal processes, or macro conditions can change both recovery amounts and timing.

References

IFRS 9 Financial Instruments — IFRS Foundation
Guidelines on PD estimation, LGD estimation and treatment of defaulted assets — European Banking Authority
Guidelines on PD estimation, LGD estimation and treatment of defaulted assets (EBA-GL-2017-16) — European Banking Authority
Basel Framework: IRB approach and credit-risk components — Basel Committee on Banking Supervision
Guidance on credit risk and accounting for expected credit losses — Basel Committee on Banking Supervision

Educational material only. Methods and examples are simplified; state and validate the assumptions required for your use case.