Loss Given Default (LGD) is a credit risk metric representing the percentage of a loan’s outstanding balance that a lender expects to lose in the event of borrower default, after accounting for all expected recoveries from collateral liquidation, collections activity, and other sources — serving as one of the three core components of credit risk quantification alongside Probability of Default (PD) and Exposure at Default (EAD) in both internal risk management and regulatory capital frameworks.
Introduction to Loss Given Default (LGD)
LGD captures the severity of a credit loss — not whether a loan will default (that is Probability of Default), but how much of the outstanding balance will be lost when it does. A loan with a 15% probability of default but 90% LGD (a typical unsecured consumer loan that defaults) poses very different economics than a loan with a 15% PD but 20% LGD (a well-secured auto loan where collateral recovery limits losses). Understanding LGD — and quantifying it accurately — is essential for credit pricing, loss reserve adequacy, and the economic sustainability of a lending business across credit cycles.
LGD varies substantially across loan product types, primarily driven by collateral type and liquidation value. Unsecured consumer loans — personal installment loans, credit cards — typically have high LGD (60-90%) because there is no collateral to recover; recoveries come only from post-charge-off collections, which are often limited. Secured auto loans have lower LGD (20-50%) because the vehicle can be repossessed and sold, though depreciation, auction costs, and condition uncertainty mean recoveries are rarely close to the outstanding balance. Residential mortgage loans can have LGD ranging widely (10-60%+) depending on property value, location, and economic conditions at the time of default. For context on regulatory capital treatment of credit risk, see Federal Reserve research on LGD estimation.
How Loss Given Default (LGD) Works
LGD is calculated as: LGD = 1 – Recovery Rate, where Recovery Rate = (Total Recoveries) / (Exposure at Default). Total recoveries include all cash flows received after default: proceeds from collateral liquidation (net of repossession costs, storage, auction fees, and sales commissions), post-charge-off collections payments (whether collected in-house or by third-party agencies), proceeds from debt sales (when the charged-off loan is sold to a debt buyer at a fraction of face value), and insurance proceeds (for credit life, credit disability, or GAP coverage). The recovery rate calculation is time-sensitive — a recovery of $2,000 on a $5,000 charged-off loan in month three post-charge-off is worth more in present value than the same recovery in month 24, so LGD calculations for pricing and capital purposes typically use discounted recoveries.
Empirical LGD estimation requires historical data on defaulted loans — what the outstanding balance was at default, what recoveries were realized, and over what timeline. This data must be segmented by product type, collateral type, vintage, and other relevant characteristics because LGD can vary significantly across segments. A lender with five years of charge-off and recovery history has the data needed to build reliable LGD estimates for its existing products; a lender launching a new product type must rely on industry benchmarks or proxy data until sufficient in-house experience accumulates. LGD estimates should be updated regularly as the recovery environment changes — collateral values, collection effectiveness, and debt buyer bid prices all fluctuate with economic conditions, and LGD assumptions that were accurate in a strong economy may materially understate losses in a recession.
LGD is used in the Current Expected Credit Loss (CECL) model that all U.S. lenders must now use for allowance for loan and lease losses (ALLL) estimation. Under CECL, the allowance is the product of PD × LGD × EAD, summed across the portfolio, adjusted for the reasonable and supportable economic forecast. Accurate LGD estimation is therefore directly tied to allowance adequacy — understating LGD understates reserves and overstates reported net income; overstating LGD overstates reserves and understates reported income. Both errors have regulatory and financial reporting consequences. See FDIC charge-off and recovery data for industry benchmarking of LGD by product type.
Example
A consumer auto title lender analyzes three years of charge-off data on 840 defaulted loans to develop its LGD estimate for reserve setting. The analysis shows: average balance at charge-off of $2,340; average vehicle liquidation proceeds net of repossession costs of $1,180 (50% of the charged-off balance); average post-liquidation collections of $120 (5% of the charged-off balance); average debt buyer proceeds on unsold deficiency balances of $85 (6% of the remaining balance after liquidation). Total average recovery: $1,385, or 59% of the charged-off balance, yielding an LGD of 41%. The lender uses this 41% LGD estimate, combined with its 8.2% estimated portfolio PD and average EAD equal to 92% of current outstanding balance, to calculate a CECL allowance of approximately 3.1% of the portfolio — an increase from the prior incurred-loss reserve of 1.9%. The difference represents the CECL transition adjustment, which must be recorded through equity upon adoption.
LGD and Loan Pricing
LGD is a fundamental input to risk-based loan pricing. The minimum yield a lender must charge on a loan to cover its expected loss is a function of PD × LGD — higher expected losses require higher yield to generate the same risk-adjusted return. A lender pricing an unsecured personal loan with 15% PD and 80% LGD needs to cover 12 percentage points of expected loss annually, requiring a yield well above that level before accounting for operating costs and desired profit margin. By contrast, a secured auto loan with 8% PD and 35% LGD requires only 2.8 percentage points of expected loss coverage — a dramatically lower loss hurdle that allows competitive pricing at much lower rates.
Understanding the LGD dynamics of each product type enables lenders to design products with appropriate collateral requirements, origination standards, and pricing that generate sustainable risk-adjusted returns. Lenders who underestimate LGD — pricing as if recoveries will be 60% when actual recoveries are 30% — systematically underprice their loans and will eventually face capital adequacy issues as actual losses exceed projections. The discipline of rigorous LGD estimation and regular back-testing against actual recovery experience is therefore a financial survival requirement for lending businesses with significant charge-off exposure.
Bottom Line
LGD is a foundational metric for loan pricing, reserve adequacy, and portfolio risk management — and inaccurate LGD assumptions lead directly to mispriced loans, inadequate reserves, and regulatory capital deficiencies. Vergent LMS provides the loan-level transaction data and charge-off history tracking needed to support rigorous LGD analysis, including complete recovery tracking after charge-off and collections management data that enables lenders to calculate actual recovery rates with the granularity required for CECL modeling and credit policy decisions.
