The Analysis of Banking Crisis Determinants

This work discusses the phenomenon of banking crisis and explains the concept of its possible prediction based on the values of various macroeconomic and institutional indicators via different methods. The study takes into account crises in the period from 1980 to 2012. One of the approaches consists in employing a panel and non-panel logit regressions, determining significant indicators and making judgments on the basis of group significance of indications. Another approach (the signal one) deals with different indicators and the extent of their predictive power for forecasting a crisis is determined through hitting certain levels and occurrence of a crisis afterwards. The logit approach made clear the joint significance of the following indicators: broad money to total reserves ratio, deposit interest rate, short term debt of total reserves, the presence of deposit insurance, total reserves of total external debt, reserves and related items, real effective exchange rate index and GDP per capita growth (annual). The signal approach yielded the significance of the following indicators: broad money to total reserves ratio, deposit interest rate, short term debt of total reserves, total reserves of total external debt, reserves and related items, real effective exchange rate index and GDP per capita growth (annual). The presence of deposit insurance was represented as a dummy variable in the logit model, so its significance could not be verified via the method of signals.