Dating, detecting and forecasting financial crises in the EU accession countries

Abstract

The recent global financial crisis has rekindled the interest of economists and policymakers in forecast modelling. In this paper we try to answer the following two questions: 1. Can an early warning system (EWS) model predict the occurrence of a financial crisis in a specified time horizon in the EU candidate countries? and 2.What are the best performing predictors of financial crisis in this group of countries? By specifying a binomial logit model based on actual quarterly panel data for the four EU candidate countries (Croatia, Iceland, Macedonia and Turkey) over the sample period of January 2005 (2005 Q1) to June 2010 (2010 Q2), we find the following four explanatory variables: the GDP growth rate, the trade balance as a percentage of GDP, the ratio of bank deposits to GDP and the budget balance as a percentage of GDP the best performing early warning indicators of financial crises incidence. Furthermore, we find that the estimated binomial logit model has a strong predictive power. Based on that fact, we conclude that early warning indicators of financial crisis do work.