Predicting Early Warning Signals of Financial Distress: Theory and Empirical Evidence

Abstract

This study proposes a simple theoretical framework that allows for assessing financial distress up to five years in advance. We jointly model financial distress by using two of its key driving factors: declining cash-generating ability and insufficient liquidity reserves. The model is based on stochastic processes and incorporates firm-level and industry-sector developments. A large-scale empirical implementation for US-listed firms over the period of 1980-2010 shows important improvements in the discriminatory accuracy and demonstrates incremental information content beyond state-of-the-art accounting and market-based prediction models. Consequently, this study might provide important ex ante warning signals for investors, regulators and practitioners.

Keywords: Financial distress prediction, probability of default, accounting information, stochastic processes, simulation

JEL Classification: C63, C52, C53, G33, M41