Deriving the term-structure of loan write-off risk under IFRS 9 by using survival analysis: A benchmark study

2026-03-12

AI Summary

This study addresses a key challenge in Loss Given Default (LGD) modeling under the IFRS 9 framework: accurately estimating time-varying write-off probabilities post-default to construct the term structure of write-off risk. It is the first to systematically apply survival analysis methods—specifically discrete-time hazard models and conditional inference survival trees—within a two-stage LGD modeling framework to estimate marginal write-off probabilities over the default horizon. The paper innovatively introduces a binarization step that converts these estimated probabilities into binary (0/1) variables for subsequent LGD computation. Empirical results show that the discrete-time hazard model outperforms other two-stage approaches across most diagnostic metrics; however, single-stage models exhibit superior overall performance, likely due to the L-shaped distribution of LGD data. This work provides a reproducible benchmark for LGD modeling.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Reasoning under Uncertainty: Relational Probabilistic Models
• Natural Language Processing: (Large) Language Models

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

The estimation of marginal loan write-off probabilities is a non-trivial task when modelling the loss given default (LGD) risk parameter in credit risk. We explore two types of survival models in estimating the overall write-off probability over default spell time, where these probabilities form the term-structure of write-off risk in aggregate. These survival models include a discrete-time hazard (DtH) model and a conditional inference survival tree. Both models are compared to a cross-sectional logistic regression model for write-off risk. All of these (first-stage) models are then ensconced in a broader two-stage LGD-modelling approach, wherein a loss severity model is estimated in the second stage. In expanding the model suite, a novel dichotomisation step is introduced for collapsing the write-off probability into a 0/1-value, prior to LGD-calculation. A benchmark study is subsequently conducted amongst the resulting LGD-models. We find that the DtH-model outperforms other two-stage LGD-models admirably across most diagnostics. However, a single-stage LGD-model still had the best results, likely due to the peculiar `L-shaped' LGD-distribution in our data. Ultimately, we believe that our tutorial-style work can enhance LGD-modelling practices when estimating the expected credit loss under IFRS 9.

Problem

• loan write-off risk
• term-structure
• LGD
• IFRS 9
• credit risk

Innovation

• survival analysis
• term-structure of write-off risk
• two-stage LGD modeling
• discrete-time hazard model
• IFRS 9