Loss Given Default (LGD) KPI

Loss Given Default (LGD) is a critical performance indicator that quantifies potential losses when a borrower defaults.

It directly influences financial health, risk management strategies, and capital allocation decisions.

A lower LGD indicates effective credit risk management, while a higher value may signal vulnerabilities in the lending process.

Organizations that optimize LGD can enhance their forecasting accuracy and improve their overall ROI metric.

By understanding this KPI, executives can make data-driven decisions that align with strategic goals and drive better business outcomes.

How Loss Given Default (LGD) Connects to Your Strategy

Within the Financial Risk Management KPI group, Loss Given Default ranks thirteenth by priority. It sits just ahead of Exposure at Default and well behind the co-metrics customers open the group with: Capital Adequacy Ratio at the top, then Credit Risk, Risk-Adjusted Return on Capital, Value at Risk, and Stress Testing. LGD carries a financial placement on the balanced scorecard, in keeping with a group that is mostly financial in perspective. A financial placement here means the metric is about money recovered and money lost, not about the workflow that produces those outcomes.

LGD is the severity parameter in the Basel internal ratings based framework, one of three that combine to estimate expected loss with Probability of Default and Exposure at Default. Read the three as a sequence of questions. Probability of Default asks whether a borrower fails. Exposure at Default asks how much is outstanding when they do. LGD asks the last and often hardest question: once you have chased down collateral and worked through recovery, what share of that exposure is gone for good? The answer turns on recovery assumptions, and small shifts in those assumptions move expected loss and the capital held against it.

What drives LGD is collateral and the realistic value you can recover from it, net of the time and cost of the workout. That is where the tension lives. Refined severity estimates feed Credit Risk segmentation, letting teams separate well-secured facilities from thin ones, but the same estimates shape Risk-Adjusted Return on Capital: assume optimistic recoveries and a facility looks more profitable per unit of capital than it is, while conservative recoveries tie up capital that the business would rather deploy. So LGD pulls between prudence in the risk function and return targets in the front office, and the collateral assumptions sit right at that seam.

Measuring Loss Given Default (LGD) in Practice

LGD data is scattered across the systems that record what actually happened after a default. Recovery and workout records hold the cash collected and the cost of collecting it. Collateral registers hold what was pledged and its valuation history. Loan and facility systems hold the exposure that severity is measured against. Assembling a clean loss history means stitching these together for each defaulted facility over its full resolution.

Decide these forks before you measure:

• Economic versus accounting loss. Accounting write-offs follow provisioning rules; economic loss tracks real cash in and out. They diverge, and you have to pick which one the metric represents.
• Discounting of recoveries. Recoveries arrive over months or years, so choose whether to discount them to the default date and at what rate. Undiscounted recoveries flatter severity.
• Cure treatment. Some defaults cure without loss. Whether you include cures, and how, changes the severity distribution substantially.
• Downturn calibration. Deciding whether to condition estimates on stressed conditions, and which period counts as the downturn, shapes the figure used for capital.

Segment where the drivers actually differ. Seniority and collateral type are the strongest, so separate senior secured from unsecured, and split collateral by type rather than pooling it. Asset class matters too, since a corporate loan book and a retail book recover on different paths.

Guard against instrumentation traps. Survivorship bias creeps in when facilities still in workout are dropped from the sample, since the slow, painful cases are often the worst ones. Incomplete workout windows understate loss when a resolution is counted before it finishes. And mixing a regulatory severity definition with a management one in a single pull produces a blended figure that means neither.

Common Pitfalls

Many organizations misinterpret LGD as a static figure, overlooking its dynamic nature influenced by market conditions and borrower behavior.

• Failing to regularly update recovery models can lead to outdated assumptions. This results in inflated LGD values that do not reflect current realities, skewing risk assessments.
• Neglecting to segment borrowers based on creditworthiness distorts overall LGD calculations. Without this granularity, organizations may misallocate resources and misjudge risk exposure.
• Ignoring macroeconomic indicators can lead to misguided confidence in LGD estimates. Economic downturns can significantly increase default rates, necessitating a reevaluation of risk models.
• Overlooking the importance of recovery strategies can inflate LGD figures. Organizations must actively manage collections and asset recovery to minimize losses effectively.

Improvement Levers

Enhancing LGD requires a proactive approach to credit risk management and recovery processes.

• Implement advanced analytics to refine credit scoring models. Data-driven insights can help identify high-risk borrowers earlier, allowing for timely interventions.
• Regularly review and update recovery strategies to align with changing market conditions. Adapting approaches based on past performance can improve recovery rates and lower LGD.
• Invest in training for collections teams to enhance negotiation skills. Well-trained staff can increase recovery rates, directly impacting LGD figures.
• Utilize technology to streamline recovery processes. Automation can reduce manual errors and improve efficiency, leading to faster resolution of defaults.

Loss Given Default (LGD) Benchmarks

We have 5 relevant benchmarks in our benchmarks database.

Browse the Top Benchmarked KPIs in Financial Risk Management

Reading the Benchmarks for Loss Given Default (LGD)

The sources cited for Loss Given Default share a name but measure genuinely different constructs, which is why a loose figure taken from one is meaningless against another. The divergence is in the definition itself, not only in the sample.

• The Federal Reserve reference is scenario based. It produces severity under a hypothetical stress path for corporate loans rather than from a bank's realized experience, so it describes what loss could look like in a prescribed adverse world, not what has been observed.
• The Moody's reference covers defaulted structured finance securities. Severity there is closer to a market based measure, read from the price of the instrument after default, and structured finance securities behave nothing like whole loans, so the construct differs at the root.
• Global Credit Data pools realized loss experience on defaulted loans to large corporates, contributed by member banks across markets. This is workout or economic severity built from actual recoveries over real resolution periods, a different animal from a stressed projection or a market price.
• The Basel Committee on Banking Supervision reference is a regulatory threshold rather than an observation at all. It sets a supervisory reference for severity used in capital calculations, which anchors the regulatory floor rather than reporting what happened in any portfolio.

Several methodological forks separate these constructs. Workout or economic severity is measured from cash recovered over a resolution period; market severity is read from post-default price; downturn severity conditions the estimate on stressed conditions. Whether recoveries are discounted back to the default date, and at what rate, changes the result before any comparison begins. Seniority and collateral drive severity hard, and a senior secured position is not comparable to a subordinated one. Asset class matters just as much, since corporate loans and structured finance securities recover through entirely different mechanisms. And a regulatory floor is a construct set by supervisors, not a number drawn from data. Put together, these differences mean a severity figure is only interpretable once its definition, its discounting, its seniority, and its asset class are stated. A free number without that context tells a customer nothing.

OKRs That Use Loss Given Default (LGD)

Loss Given Default appears directly in this KPI group's own best practice material, so its objective grounding is explicit rather than inferred. The group records the objective Strengthen capital resilience to absorb financial shocks and maintain regulatory compliance, and its documented best practice calls in plain terms to integrate Probability of Default and Loss Given Default into credit risk segmentation. That best practice is the group's genuine, stated way of putting LGD to work: use severity estimates to sort borrowers into sharper risk categories, spot the thinly secured portfolios early, and respond with tighter collateral demands or amended covenants.

Linking that practice to the objective is straightforward. Better severity estimates feed cleaner segmentation, cleaner segmentation improves the loss and capital picture, and a sounder capital picture is what strengthening capital resilience means in practice.

Keep key results directional rather than tied to a level, and never borrow a numeric target from a neighboring key result. Reasonable framings include lifting the share of defaulted facilities with a completed workout history feeding the estimate, embedding severity bands into the segmentation model so high-severity portfolios surface earlier, or reconciling economic and regulatory severity definitions so the same facility is not measured two ways. Each supports the capital resilience objective through the group's own segmentation practice, without inventing a target the group never set.

See OKR Examples for Financial Risk Management

KPI Categories

This KPI is associated with the following categories and industries in our KPI database: