Predicting Financial Crisis With Artifıcial Neural Network Model: An Application For Turkey

Abstract

Financial crises have become a topic that has been discussed more in the post-1980 period. Capitalism, by its nature, tends to lead to a crisis. Therefore, many crises have occurred in the historical process. The interest in financial crises increased after 1980. The main reason for this is the increase the effect of financial crises and their effects of global. Financial crises have sociological effects besides their destructive effects on the country's economy. This feature of crises has lead the people working on this problem to find the factors that cause crises. One of the main issues in creating new financial architecture is the determination of financial crisis indicators. Many different methods are used to identify crisis indicators. These models are commonly referred to as the early warning system (EWS). The main reason that pushes economists to work on this issue is the foreign exchange loss, excessive recession and negative growth rates created by financial crises in the countries. Therefore, it will be of great benefit to define reliable indicators that will predict financial crises. Although the factors that cause crises are known, financial crises are reappearing. In studies related to the predictability of financial crises, it is aimed to identify and monitor the variables likely to cause problems in the future. In this context the study, data from the 2008 crisis in Turkey will be used. Which indicators should be monitored for the next crisis will be tried to be determined by using artificial neural network model.