Files
Abstract
My dissertation has two essays on the information value of insurer rating changes. The first essay conducts the event studies of rating changes for public insurers. Using a carefully designed benchmark portfolio model, I find that the market has a significant short-run negative response to downgrades, but no significant response to upgrades, consistent with the asymmetric response evidence in the existing literature. I also find a post-event drift during the 3-month period following the downgrades but not the upgrades, and a pre-event market adjustment during the 30-day and 12-month period prior to rating changes for both downgrades and upgrades. These results are consistent with the incomplete market response hypothesis and rating change anticipation hypothesis. Using the subsample analysis, I test 12 relevant hypotheses, and find that both short and long run market response to downgrades are stronger for small insurers, low-credit-quality insurers, and for subsequent downgrades, mono-line firm downgrades, and threshold downgrades. The second essay employs the Granger-causality to study the causality relationships among insurer rating changes, stock returns and insurer performances measured by return on equity. I find that excess stock returns are generally good predictors of future firm performance, while rating changes are not. Furthermore, rating changes can be predicted by past excess stock returns and risk adjusted return on equity.