Paper Title
Do Early Warning Signals Predict Financial Crises? Evidence From Asian Countries

Abstract
The Asian crisis of 1997 was seen as one of the most terrific dangers to Asian emerging countries. This crisis was driven, among others, by the massive indebtedness of some countries that have over-invested in the 90s in unprofitable or very risky projects, particularly in the real estate sector. An important part of the private sector debt was contracted in foreign currencies (mainly in US dollars) without hedging. As currency risk was significant, a depreciation of the domestic currency against the US dollar increases mechanically the amount of debts. Therefore, these market imperfections weakened Asian banks which suffered deterioration in their balance sheets due to losses and the increase in impaired loans. This situation ended up with devaluation of local currencies which lead to the emergence of the crisis. Faced with the recurrence of banking crises in recent decades, several regulatory authorities have developed some models in the line of early warning systems (EWS) in order to predict crises. These models of crises prediction enable decision-makers to identify economic weaknesses and vulnerabilities in order to take preventive measures to reduce the risk of appearance of a crisis. The aim of our paper was to determine the variables that may have a role in predicting banking crisis in 6 Asian emerging countries over the period 1973-2012. Macroeconomic, external and financial indicators of banking crisis were selected. Based on panel logit model, our results reveal that inflation has the strongest impact in predicting systemic banking crisis, while economic growth, real exchange rate, the ratios M2 by foreign exchange reserves, credit to private sector by GDP and short-term debt to external debt have a small effect. Keywords� Banking Crises, Early Warning System, Panel Logit Regression.