Using market variables in financial distress prediction for Vietnamese listed companies

The topic of financial distress prediction has demonstrated its popularity in the existing literature on the field of corporate finance because of the usefulness of a financial prediction model in risk management for not only companies’ managers but also investors and creditors. This paper aims to investigate the classification power of market variables as predictors in the financial distress prediction model for listed companies in a frontier market as Vietnam securities market. Data is collected from 70 financially distressed companies that suffer a loss in 3 consecutive years and 156 non-financially distressed companies in Vietnam from 2010 to 2017. Four different models have been constructed using Logit regression and SVM analysis technique to make a prediction in 1 to 3-year ahead. The analysis results show that combining accounting ratios with market variables such as price volatility and P/E can improve the classification ability of the ex-ante model. In addition, contrary to the results of related previous researches in emerging markets, in this study, Logit models outperform SVM models. Therefore, for future research, models that apply other machine learning classifiers such as Decision Tree (DT) or Neural Network (NN) should be investigated.