Early Warning System for Failing Nations

This study addresses the need for an early warning system of financial distress of nations, amply demonstrated by failures in Mexico and South East Asia. We find that financial ratios of failing nations, and those of failing corporations, display similar patterns. A neural network was trained on financial ratio-observations from defaulting and non-defaulting debtor nations 1989-1993. The model's output is a 1994-1996 index, separating 220 developing nations into high and low risk default categories with 77% accuracy, as measured by Paris Club incidents among developing nations in subsequent years. Ratios relative to liquidity, ability to pay, and economic activity were confirmed to have predictive qualities; ratios relative to credit worthiness were not confirmed to have predictive qualities. IMF/World Bank policies for the past three decades, plus future research ideas, are presented