Comparison between Altman Z-Score and Machine Learning in Predicting Financial Distress of Energy Companies in Indonesia

Company financial health is a crucial aspect that affects business continuity, capital market stability, and investor confidence. Financial distress poses a serious threat to companies, especially in developing countries like Indonesia, as it can lead to defaults and bankruptcy with broad impacts. This situation worsened due to multidimensional crises, including the COVID-19 pandemic (2020–2022), which caused revenue declines and rising debt burdens across sectors, prompting the need for reliable early detection systems. This study aims to build a financial distress prediction model for Indonesian companies, particularly in the energy sector, during 2020–2023, using machine learning methods such as Logistic Regression, XGBoost, and Random Forest. Additionally, the study compares the performance of these machine learning models with traditional approaches like Altman Z-Score (1968) and Zmijewski X-Score (1984) to assess prediction accuracy. Findings in this study prove that machine learning methods, especially XGBoost, provide more accurate prediction in identifying financial distress. This predictive model functions as an early warning system beneficial for investors’ decision-making, corporate management financial strategy improvement, and regulators to maintain national capital market stability. Thus, financial distress prediction not only mitigates bankruptcy risk but also serves as a strategic tool for fostering healthier and more sustainable corporate governance in Indonesia’s energy sector.