Prognostic Power of Early Warning Signals for Financial Crises – Theoretical Approaches and Empirical Results

The accurate forecasting of global maladies and catastrophes has been an area of paramount importance in many fields of study. Over the last years, the interest in financial crises has experienced a definite resurgence in academic spheres, an interest which had largely dissipated by the end of the twentieth century. This was largely the consequence of decades of high global growth rates, growing international trade and capital flows, accompanied by low inflation and low interest rates (the so-called Great Moderation). The specter of financial crises had all but disappeared, but the Global Financial Crisis, beginning in 2007, was quick to remind academics, politicians, and businessmen of the multifarious mechanisms through which crises can develop, as well as the grave economic and social repercussions that follow. Understandably, this rekindled the interest in “early-warning” models of such crises, and the debate over their validity and forecasting abilities was quickly reestablished. This article modifies the well-known KLR model and tests its applicability for Bulgaria. Using the large quantity of empirical data available, it makes specific calculations for the behavior of a group of selected indicators for the early prediction of crises.