Credit Ratings and Predictability of Stock Returns and Volatility of the BRICS and the PIIGS: Evidence from a Nonparametric Causality-in-Quantiles Approach

This paper provides a novel perspective to the predictive ability of credit rating announcements over stock market returns and volatility using a novel methodology that formally distinguishes between different market states that can be characterized as bull, bear and normal market conditions. Using data on the credit rating announcements published by the three well-established credit rating agencies and data on BRICS and PIIGS stock markets, we show that the stock markets react heterogeneously, and in quantile-specific patterns, to ratings announcements with more persistent and widespread effects observed for PIIGS stock markets. The effect of rating announcements is generally stronger and more widespread in the case of volatility of returns, implying significant risk effects of these announcements. Finally, we show that the results of the aggregate ratings are driven mostly by rating upgrades rather than downgrades, implying asymmetry in the predictive ability of ratings announcements during good and bad times. Overall, our findings show that predictive models can be greatly enhanced by disaggregating the overall rating announcements and taking into account nonlinearity in the relationship between ratings announcements and stock return dynamics.