Explaining the Model of Bankruptcy Prediction to Identify Healthy and Risky

Bank failure is a phenomenon that has attracted the attention of the central bank, banks and financial institutions. Since the signs of potential failure are detected before the bankruptcy, identifying warning variables, and prediction of the crisis timely provide an opportunity for managers and creditors to create preventive activities. In this paper, using the financial statements in the period 1385-1393 and Z-score index as an indicator of bankruptcy, bank failures are detected. To identify the failed banks, we use the kernel function of the index and the stress of index is calculated. Banks that are under stress point are considered risky and otherwise healthy. For estimate of model, discriminate analysis was used to identify the factors that enable to distinguish healthy and risky banks, then, using the logit model, a model was developed for predicting banks failure. To verify separation of two groups of banks we checked Wilks Lambda, F and two independent samples mean, then to evaluate the importance of the different independent variables in a model we used intercorrelation between variables. The results showed 87 percent accuracy of discriminant analysis and 98.2 percent logit model in compliance with the Iran`s banking network environment.