Credit Risk Management for Mongolian Banks Essay

The importance of optimal decision-making and precise predictions is not limited to banks only but also of importance to other financial institutions. Nowadays, financial markets are becoming increasingly uncertain and interdependent, making accurate prediction of future market directions a near impossible task. Although, in case of Mongolia, some econometric models are being tested for the last two decades, practical application is lackluster and it is common practice for businessmen to make decisions based on intuition and gut feeling. Unfortunately, this unscientific approach to decision making is quite commonplace. The objective of this research is to overcome conditions and to identify the best evaluation model for credit risk forecasting for banking institutions. From a theoretical point of view, this research paper introduces a literature review on the application of back propagation algorithm of an artificial neural network, linear probability model, and binary choice (logit probit) model for credit risk management. Whereas, from an empirical point of view, this research compares the econometric models and artificial neural network using Mongolian banks’ credit risk data, and shows the differences between the aforementioned four models. We demonstrate that artificial neural network model is more convenient for Mongolian banks’ credit risk management than other econometric models due to the models’ evaluation and forecast accuracy. Therefore, we recommend Mongolian banks and financial institutions to apply ANN model to forecast credit risk and to hedge risk. Key words: credit risk management, linear probability model, binary choice logit and probit model, artificial neural network, back propagation algorithm, forecasting.