An Early Warning System for Identifying Financial Instability

Abstract

Financial crises prediction is an essential topic in finance. Designing an efficient Early Warning System (EWS) can help prevent catastrophic losses resulting from financial crises. We propose an EWS for predicting potential market instability conditions under which perturbations in the price level may evolve in large price declines or changes in general. Our system is based on the so called price-volatility feedback rate, which is supposed to describe the ease of the market in absorbing small price perturbations. A logit regression EWS is employed to predict future large price losses and Early Warning Indicators (EWIs) based on the realized variance (RV) and on the price-volatility feedback rate are considered. Our study reveals that, while the RV may sometimes fail in predicting future price losses in a given time series, the EWI employing the price-volatility feedback rate is always an important predictor of financial instability.