Hot Off the Press: News-implied Sovereign Default Risk

Abstract

We develop a novel sovereign default risk index using natural language processing techniques and 10 million news articles covering over 100 countries. The index is a reliable high-frequency measure of countries' default risks, particularly for those lacking market-based measures. It co-moves with and predicts sovereign CDS spread, and predicts sovereign credit rating downgrades and realized sovereign defaults over long horizons. Consistent with theory on sovereign default risk spillovers, an increase in the index predicts (i) higher firm default probability, (ii) higher cost of protection against firm default, and (iii) lower equity market valuation, which is much stronger for global, compared to country-specific, default concerns. Overall, the index yields valuable signals for profitable equity market timing strategies.

Keywords: Sovereign default, Credit risk, Equity returns, Machine learning, Natural language processing, Early warning indicators

JEL Classification: F10, F30, G12, G15