Abstract:There has been intensive research from academics and practitioners regarding models for predicting bankruptcy and default events, for credit risk management. Seminal academic research has evaluated bankruptcy using traditional statistics techniques (e.g. discriminant analysis and logistic regression) and early artificial intelligence models (e.g. artificial neural networks). In this study, we test machine learning models (support vector machines, bagging, boosting, and random forest) to predict bankruptcy one year prior to the event, and compare their performance with results from discriminant analysis, logistic regression, and neural networks. We use data from 1985 to 2013 on North American firms, integrating information from the Salomon Center database and Compustat, analysing more than 10,000 firm-year observations. The key insight of the study is a substantial improvement in prediction accuracy using machine learning techniques especially when, in addition to the original Altman’s Z-score variables, we include six complementary financial indicators. Based on Carton and Hofer (2006), we use new variables, such as the operating margin, change in return-on-equity, change in price-to-book, and growth measures related to assets, sales, and number of employees, as predictive variables. Machine learning models show, on average, approximately 10% more accuracy in relation to traditional models. Comparing the best models, with all predictive variables, the machine learning technique related to random forest led to 87% accuracy, whereas logistic regression and linear discriminant analysis led to 69% and 50% accuracy, respectively, in the testing sample. We find that bagging, boosting, and random forest models outperform the others techniques, and that all prediction accuracy in the testing sample improves when the additional variables are included. Our research adds to the discussion of the continuing debate about superiority of computational methods over statistical techniques such as in Tsai, Hsu, and Yen (2014) and Yeh, Chi, and Lin (2014). In particular, for machine learning mechanisms, we do not find SVM to lead to higher accuracy rates than other models. This result contradicts outcomes from Danenas and Garsva (2015) and Cleofas-Sanchez, Garcia, Marques, and Senchez (2016), but corroborates, for instance, Wang, Ma, and Yang (2014), Liang, Lu, Tsai, and Shih (2016), and Cano et al. (2017). Our study supports the applicability of the expert systems by practitioners as in Heo and Yang (2014), Kim, Kang, and Kim (2015) and Xiao, Xiao, and Wang (2016).

Predicting financial crises: an evaluation of machine learning algorithms and model explainability for early warning systems

Assessing Machine Learning Techniques for Predicting Banking Crises in India

Understanding the Great Recession Using Machine Learning Algorithms

Machine learning models and bankruptcy prediction

Developing an Early Warning System for Financial Networks: An Explainable Machine Learning Approach

Can Text-Based Statistical Models Reveal Impending Banking Crises?

Analysis of Financial Credit Risk Using Machine Learning

Improved SVM Algorithm Financial Management Model for Data Mining

Multinomial logit as an early warning model for predicting banking crises

CRISIS ALERT:Forecasting Stock Market Crisis Events Using Machine Learning Methods

Assessing machine learning for forecasting economic risk: Evidence from an expanded Chinese financial information set

Machine Learning-Based Financial Statement Analysis

Predicting Economic Recessions Using Machine Learning Algorithms

Explainable Machine Learning for Financial Distress Prediction: Evidence from Vietnam

Can Machine Learning Catch Economic Recessions Using Economic and Market Sentiments?

Early warning of corporate financial crisis based on sentiment analysis and AutoML

Large-scale data-driven financial risk management & analysis using machine learning strategies

Lost in noise? Some thoughts on the use of machine learning in financial market risk measurement

Comparative Analysis of Machine Learning, Hybrid, and Deep Learning Forecasting Models Evidence from European Financial Markets and Bitcoins

Machine Learning for an Enhanced Credit Risk Analysis: A Comparative Study of Loan Approval Prediction Models Integrating Mental Health Data