The impact of class imbalance in logistic regression models for low-default portfolios in credit risk

Paper: arXiv 2602.19663

Authors: Willem D. Schutte, Charl Pretorius, Neill Smit, Leandra van der Merwe, Robert Maxwell

Abstract

In this paper, we study how class imbalance, typical of low-default credit portfolios, affects the performance of logistic regression models. Using a simulation study with controlled data-generating mechanisms, we vary (i) the level of class imbalance and (ii) the strength of association between the predictors and the response. The results show that, for a given strength of association, achievable classification accuracy deteriorates markedly as the event rate decreases, and the optimal classification cut-off shifts with the level of imbalance. In contrast, the Gini coefficient is comparatively stable with respect to class imbalance once sample sizes are sufficiently large, even when classification accuracy is strongly affected. As a practical guideline, we summarise attainable classification performance as a function of the event rate and strength of association between the predictors and the response.

Complexity vs Empirical Score

• Math Complexity: 4.5/10
• Empirical Rigor: 7.5/10
• Quadrant: Street Traders — practical and empirical, lighter on theory

Why this score: The paper provides a well-structured simulation study analyzing the effects of low event rates and signal strength on logistic regression models in credit risk. While mathematically straightforward (relying on standard WoE and logistic regression formulations), it offers high practical rigor and clarity for financial risk modeling.

Research Flowchart

  flowchart TD
    A["Research Question: Impact of class imbalance on <br>logistic regression in low-default credit portfolios"] --> B["Simulation Study Setup"]
    
    B --> C["Inputs & Varied Parameters:<br>- Event Rates (Class Imbalance)<br>- Predictor Association Strengths"]
    
    C --> D["Computational Process:<br>Logistic Regression Modeling & Evaluation"]
    
    D --> E["Key Finding 1: Classification Accuracy<br>drops & optimal cut-off shifts with low event rates"]
    D --> F["Key Finding 2: Gini Coefficient<br>remains stable across imbalance levels in large samples"]
    
    E --> G["Outcome: Practical Guidelines for <br>Attainable Performance vs. Event Rate"]
    F --> G