TRENDS IN CORPORATE BANKRUPTCY PREDICTION MODELS: A LITERATURE REVIEW

Abstract

This study investigates the evolving landscape of corporate bankruptcy prediction models in Indonesia, focusing on their methodological diversity, sectoral applicability, and contextual relevance in the post-pandemic era. The increasing financial fragility of Indonesian firms particularly in manufacturing, construction, and services underscores the urgency for accurate early-warning systems tailored to local economic dynamics. This review fills a gap in prior research by systematically mapping traditional models (e.g., Altman Z-Score, Zmijewski, Ohlson) and emerging approaches (e.g., machine learning, hybrid models) within the Indonesian context. Using a narrative literature review method, 32 relevant studies were thematically analyzed to capture trends in model usage, performance, and sector-specific calibration. The findings reveal that while classical models remain dominant in capital-intensive industries due to their simplicity and interpretability, their performance weakens in asset-light sectors. In contrast, machine learning methods—such as Random Forest and XGBoost—demonstrate higher predictive accuracy in dynamic and data-rich environments but face challenges of interpretability and regulatory acceptance. A growing body of research also supports the integration of non-financial variables, including ESG indicators, audit quality, and corporate governance. The study concludes that future research should focus on hybrid model development, underexplored sectors such as MSMEs, and improved data infrastructure. These insights offer actionable implications for policymakers, regulators, and industry stakeholders in strengthening corporate financial resilience and developing more inclusive, adaptive predictive frameworks.