Loss Given Default

What Is Loss Given Default?

Loss Given Default (LGD) is one of the three core parameters in credit risk modeling, alongside Probability of Default (PD) and Exposure at Default (EAD). It represents the fraction of a credit exposure that will not be recovered if a counterparty defaults, expressed as a percentage. Mathematically: LGD = 1 − Recovery Rate. If a bondholder recovers 40 cents on the dollar after a default, the LGD is 60%.

LGD is determined by the seniority of the debt claim, the quality and liquidity of collateral, the jurisdiction's insolvency framework, and prevailing economic conditions at the time of default. Senior secured debt in the US has historically averaged LGDs of 20–35%, while unsecured subordinated bonds can see LGDs exceeding 80%. Crucially, LGD is not a static input, it is a distribution, not a point estimate, and that distribution widens dramatically during systemic stress when correlated defaults overwhelm restructuring capacity simultaneously.

Why It Matters for Traders

For credit investors, whether trading high yield bonds, leveraged loans, collateralized loan obligations (CLOs), or sovereign CDS, LGD is just as important as default probability in assessing fair value. The expected loss formula (EL = PD × LGD × EAD) underpins spread levels across the entire credit universe. A bond trading at a 500 basis point spread might appear cheap on a PD basis alone but fairly priced or even expensive once realistic LGD assumptions are incorporated.

During credit cycle downturns, LGD tends to spike cyclically because distressed asset sales occur at depressed prices, collateral values fall, and the sheer volume of simultaneous defaults overwhelms the legal system's restructuring capacity. This wrong-way risk dynamic, where default frequency and severity rise together, means credit spreads during recessions must compensate for both higher default rates and lower recoveries simultaneously. Equity-to-credit crossover investors frequently underappreciate this compounding effect, focusing on PD while treating recovery as constant at historical averages.

In structured credit, LGD assumptions are the critical lever in CLO waterfall modeling. A modest shift from a 60% to a 70% LGD assumption on the underlying leveraged loan pool can extinguish equity tranche value entirely and begin eroding mezzanine tranches, which is why CLO equity investors spend considerable effort stress-testing recovery scenarios rather than just default frequencies.

How to Read and Interpret It

LGD benchmarks by asset class (approximate historical US averages based on Moody's and S&P long-run studies):

• Senior secured bank loans: 20–30% LGD (70–80% recovery)
• Senior unsecured bonds: 55–65% LGD
• Subordinated / junior bonds: 70–85% LGD
• Convertible bonds: 60–70% LGD (equity optionality collapses in default)
• Sovereign debt (restructuring): Highly variable, 30–80% LGD depending on political negotiation and haircut structure

Key signals that LGD is rising in a credit cycle:

1. Falling collateral valuations in leveraged loan books, commercial real estate and enterprise value multiples compress simultaneously with credit stress
2. Rising share of covenant-lite loans, by 2021–2022, over 85% of new US leveraged loan issuance was covenant-lite, stripping creditors of early intervention rights
3. Increasing distressed debt trading volumes at prices below 70 cents on the dollar, signaling the market is pricing deep impairment
4. Jurisdiction-specific legislative risk, debtor-friendly bankruptcy reforms, administrative stays on creditor actions, or sovereign debt moratorium legislation
5. Elevated debt-to-EBITDA multiples at origination, loans issued at 7x+ leverage in benign conditions offer thin enterprise value cushions when multiples contract in downturns

Historical Context

During the 2008–2009 Global Financial Crisis, average recovery rates on defaulted high-yield bonds collapsed to approximately 25 cents on the dollar, implying LGDs near 75%, compared to the long-run historical average closer to 40 cents. This represented one of the most severe recovery shortfalls on record, driven by the simultaneous nature of defaults across industries, fire-sale asset liquidations, and frozen secondary credit markets that eliminated the price discovery necessary for orderly restructurings.

In the leveraged loan market specifically, the proliferation of covenant-lite structures originated in 2006–2007 contributed materially to lower recoveries, as creditors received far less early warning to restructure positions before terminal value impairment. The episode prompted Basel III regulators to require banks to use downturn LGD estimates, stressed values calibrated to recession scenarios, rather than through-the-cycle averages, a methodology shift with significant implications for bank capital requirements.

More recently, during the COVID-19 default wave of 2020, recovery rates held somewhat better than the GFC trough, averaging closer to 35 cents for high yield, partly because central bank intervention in credit markets restored liquidity quickly, reducing the fire-sale discount. However, retail and energy sector defaults saw LGDs consistently above 70%, demonstrating that sector-specific collateral dynamics remain the dominant driver even when systemic liquidity is restored. The contrast between 2009 and 2020 outcomes is a useful reminder that LGD is as much a function of the macro and policy environment as it is of deal-level documentation.

Limitations and Caveats

LGD is notoriously difficult to estimate ex-ante. Recovery rates are highly path-dependent, they reflect economic conditions at the specific moment of default, not at origination. A loan underwritten conservatively at 4x leverage in 2019 may still generate a high LGD if the sector's enterprise value multiples have compressed 50% by the default date. Historical LGD databases suffer from sample bias, being dominated by US and European markets with well-developed Chapter 11 and administration frameworks; LGDs in emerging market jurisdictions are significantly more volatile, less predictable, and heavily influenced by political and legal factors that defy quantitative modeling.

Additionally, in distressed debt investing, an investor's own actions can materially alter their realized LGD. A distressed fund purchasing bonds at 30 cents on the dollar and then driving a restructuring that delivers 50 cents in new instruments has realized a positive return on cost, but the face-value holder recorded a 50% LGD. This creates significant ambiguity in how recovery statistics are compiled and interpreted across market participants.

Finally, LGD correlations with PD, the wrong-way risk noted above, means that using independent PD and LGD estimates in the expected loss formula systematically understates tail risk during systemic events.

What to Watch

• Covenant-lite loan issuance volumes as a leading indicator of future LGD deterioration in the leveraged loan market
• CLO equity tranche pricing and new issue spreads for implied market LGD assumptions embedded in structured credit valuations
• Moody's and S&P quarterly default and recovery studies, both publish issuer-weighted and volume-weighted recovery statistics that reveal real-time cycle positioning
• Legislative and judicial developments in major bankruptcy jurisdictions, particularly US Chapter 11 rule changes, UK Restructuring Plan case law, and any emerging market sovereign debt framework evolution
• Distressed debt ratios, the share of high yield trading below 70 cents, as a coincident indicator that market-implied LGDs are rising across the asset class