THE INFORMATION VALUE OF SOVEREIGN CREDIT RATING REPORTS
SIM KIAN JU GEOFFREY EZEKIEL
SIM KIAN JU GEOFFREY EZEKIEL
Citations
Altmetric:
Alternative Title
Abstract
We examine the information value of linguistic tone in sovereign credit rating reports issued by Moody’s Investors Service (Moody’s) using a Naïve Bayesian machine learning algorithm. We find that tone contains useful information beyond credit ratings. Tone is significantly related to abnormal sovereign credit default swap (CDS) spread changes at the time of rating change announcements, and can predict future sovereign rating downgrades. We further decompose tone into five content categories based on Moody’s sovereign rating indicators, including macroeconomic, public & external finance, debt dynamics, financial sector, and political & institutional. We find that tone related to debt dynamics in sovereign credit rating reports has greater information value than tone related to the other four types of content. We also document a systematic shift in how investors react to tone in credit rating reports after the 2009 Eurozone sovereign debt crisis. We further decompose tone into anticipated and surprise components and find that both components of negative tone have information value, whereas only surprise positive tone has information value. Overall, our study reveals the information value and content of tone in sovereign credit rating reports, and the relative importance investors place on different components of tone.
Keywords
Source Title
Publisher
Series/Report No.
Collections
Rights
Date
2014
DOI
Type
Thesis