The challenge of predicting currency crises: how do definition and probability threshold choice make a difference?

Abstract

In order to analyze the definition effect on prediction powers of Early Warning Systems (EWS), we construct an EWS model with 20 different crisis definitions. Furthermore, to analyze the probability threshold effect, we identify 11 different threshold levels and forecast the different versions of our model for each of those threshold levels. Our results show that crisis definitions and threshold choices significantly affect the prediction powers of the EWS models. To put it more explicitly, Exchange Market Pressure (EMP) Index based definition is shown to be a better predictor compared to depreciation based definition. Furthermore, EMP Index is found to give better results with higher standard deviation multiplier. Last but not least, it is empirically proven that 50 % threshold is the optimal level for EWS analyses as until that level the prediction powers of the models significantly increase but keep constant above it.