The study of corporate bankruptcy prediction models: univariate analysis and logistic regression
Corporate failure has become a major academic research over the last fifty years. During this time, various failure prediction models and bankruptcy theories were developed. From the initial univariate analysis to market-based models of the twenty-first century, the accuracy of the prediction models are improved continually and their link with corporate governance is attracting more attention. However, the emergence of new prediction methods does not mean the decline of traditional empirical analysis. To inspect the relationship between accounting ratios and bankruptcy risk, this paper investigates the efficacy of univariate analysis and logit model to predict bankruptcy problems with a sample of 52 failed and 518 non-failed companies over the period from 2004 to 2008. Further, comparing with the model incorporated with corporate governance information, including number of directors, female percentage, board average age, average tenure and outside directors percentage, I find corporate governance can enhance the accuracy of the prediction model. After combining these corporate governance variables, the accuracy of the prediction model has been improved. Based on the findings of this study, I explore the evidence of factors contributed to the bankruptcy of companies during the financial crisis.
logistic regression, the logit model, Receiver Operating Characteristics (ROC), corporate governance information
Zongyi Wang. The study of corporate bankruptcy prediction models: univariate analysis and logistic regression / Zongyi Wang // Управління розвитком складних систем : зб. наук. праць / Київ. нац. ун-т буд-ва і архітектури ; гол. ред. П. П. Лізунов. – Київ : КНУБА, 2019. – № 39. – С. 171-178. - Бібліогр. : 15 назв.