Prediction of crises in turkish economy via the early warning system.

The purpose of this study is to predict Turkish Economic Crises for the time period of 2004 to 2014 based on the "Signal Approach" developed by Kaminsky, Lizondo and Reinhart (KLR) (1998). In our study, a total of 15 indicators were tested, for the time period spanning 2004:Q1-2014:Q2. According to the findings of the study, the unemployment rate is the best-performing variable in terms of the right signal to predict crises. International reserves, real interest rate, real exchange rate, and consumer price index are the best performers in terms of the possibility to anticipate a crisis. In addition to that, they are also the best-performing variables in predicting crises in terms of signal continuity in pre-crises periods. Current account balance, reserve adequacy and public expenditure emerged as the worst performing variables in predicting crises in terms of signal and continuity. Our study demonstrated that the Composite Index of Leading Indicators, composed of all variables excluding the failing variables, is successful in predicting crises in the select period, and that the chosen leading indicators are successful as a whole in terms of predicting crises.