技術指南

預設損失與預設風險暴露

Loss given default (LGD) estimates the share of an exposure lost when a borrower defaults, while exposure at default (EAD) estimates the amount outstanding at that point.

聆聽此指南AI 語音敘述 · ~3 min

本頁閱讀時間3分鐘
  1. 概述
  2. 深入探討
  3. 現實世界的實施
  4. The Future of Loss Given Default and Exposure at Default
  5. 付諸實踐
  6. 不斷探索
  7. 常見問題

概述

They complement probability of default (PD), but each parameter answers a different question and depends on definitions, recovery assumptions and portfolio context.

深入探討

PD, LGD and EAD describe different components of credit risk. PD estimates the probability of a defined default event over a horizon. LGD estimates the loss fraction if default occurs, relative to the exposure; collateral, guarantees, recoveries, costs and timing can affect that estimate. EAD estimates the gross amount of the facility when default occurs. For an on-balance-sheet loan it relates to the drawn balance; for a revolving or off-balance-sheet facility, additional drawdowns can make EAD differ from today’s balance. A simplified expected-loss calculation is PD × LGD × EAD when the inputs are aligned to the same exposure, default definition and horizon. PD and LGD are ratios; EAD is an amount of currency under the Basel IRB framework. The product is an estimate, not a complete accounting provision or capital calculation. Regulatory capital formulas include additional conditions, and accounting standards can define expected credit loss differently. Do not confuse the simplified intuition with a bank’s official reporting method. Estimating LGD requires data on recoveries and costs after default, including how long collection takes and how collateral is valued. EAD models need data about balances and additional usage before default. Basel’s IRB requirements address representative observations, long-run experience and model validation for institutions using that approach. Changes in product terms, collections policy or economic conditions can alter the estimates. Document data, assumptions and uncertainty, and review each parameter separately before combining them in a portfolio measure.

現實世界的實施

A secured loan analyst estimates LGD using expected recovery from collateral and collection costs.

A revolving-credit model estimates EAD by accounting for possible future draws before default.

A risk team combines PD, LGD and EAD in a simplified expected-loss estimate for a portfolio.

A reviewer checks whether the default definition and recovery horizon match the portfolio data.

The Future of Loss Given Default and Exposure at Default

Credit portfolios and recovery environments change, so LGD and EAD may shift with collateral values, payment behavior and products. Stress testing and regular validation help expose where estimates depend on old conditions. New data sources can improve measurement but also introduce gaps or inconsistent definitions. Keep PD, LGD and EAD assumptions explicit and avoid presenting their product as a guaranteed loss for an individual loan. As products and recovery practices evolve, old parameters can misstate the amount exposed or recovered at default. Monitor performance by facility type and vintage, and validate assumptions after policy changes. Keep the simplified formula separate from any accounting or capital calculation required by a governing framework.

付諸實踐

為什麼這很重要

  • 成本與預算. 多年來,架構決策決定著效能和營運成本。
  • 更明確的決策. 技術教育幫助團隊選擇正確的堆疊,而不僅僅是最新的堆疊。
  • 品質管控. 更好的工程選擇可以減少生產中的可靠性事故。

留意

  • 優化一項基準測試可以隱藏更廣泛的系統弱點。
  • 基礎設施和維護成本常常被低估。
  • 隨著系統變得更加複雜,安全性和可觀察性差距可能會擴大。

後續步驟

  1. 在實施之前定義延遲、品質和成本目標。
  2. 在實際負載和資料條件下進行基準測試。
  3. 儀器監控錯誤、漂移和使用者影響。
  4. 在擴展之前準備回滾和事件回應路徑。

不斷探索

免費電子報

獲取每日 AI 簡報

每個工作日上午提供三則經過驗證的 AI 新聞,以淺顯易懂的英文撰寫。永久免費,無廣告。

每個工作日一封電子郵件。一鍵取消訂閱。我們絕不出售或分享您的地址。

測試一下自己

Take the Loss Given Default and Exposure at Default quiz

每個答案皆有即時回饋,通過課程後可獲得具備可驗證 ID 的可分享證書。

開始測驗

支持免費的 AI 教育。AI Understanding 是一個 501(c)(3) 非營利組織——絕無廣告,絕無付費牆。 捐款

常見問題

What is Loss Given Default and Exposure at Default?

Loss given default (LGD) estimates the share of an exposure lost when a borrower defaults, while exposure at default (EAD) estimates the amount outstanding at that point. They complement probability of default (PD), but each parameter answers a different question and depends on definitions, recovery assumptions and portfolio context.

What does LGD estimate?

The guide defines LGD as the loss fraction conditional on default.

What does EAD estimate?

The guide describes EAD as the amount outstanding at the time of default.

Which simplified formula illustrates expected loss in the guide?

The guide gives PD multiplied by LGD and EAD as a simplified estimate when inputs align.

Why can EAD exceed a revolving account’s current balance?

The guide notes additional draws can make revolving-credit EAD exceed today’s balance.

Which factor can affect LGD?

The guide explains LGD depends on recovery, collateral, costs and timing.