The comparative model study of Companies financial distress prediction in Iran: Probit and Backpropagation Neural Networks models

Nowadays, the financial centers compete with each other in a variable and competitive atmosphere. The quick and accurate reaction against the variable conditions of market plays important role in the condition of financial centers. These financial conditions and tightened competition arena cause the unsuccessful companies exclude from the competition arena as quick as possible. The aim of this study is to determine the patterns using the financial variables (financial ratios of profit and loss statement, and balance sheet) to enhance the ability of users of financial statements to decide to predict the financial crisis of companies. In this research, two patterns (Probity and neural networks) have been developed for financial crisis of companies for one, two, three and four years before the bankruptcy using 60 financial ratios. Then, according to the obtained results, the patterns were compared with each other and the best pattern was chosen. The statistic sample included two groups; 40 bankrupt companies which involved in article 141 of commercial code for at least three successive years and 82 non-bankrupt companies selected based on the Tobinʹs Q criteria. The data were used between 1386 and1390 and the results showed that both patterns were able to predict the financial crisis of corporations and neural network was more able in this regard.