Understanding Default Risk and Ttc Ratings Through Hazard Rates

This paper investigates the dynamics of default probabilities implied by Through-The-Cycle (TTC) ratings and default events. Using a Standard amp; Poor's database, we show that aging ereg;ects determine probability changes in the short term after a rating assignment, while the business cycle dominates in the long run. We examine to what extent ratings echo changes in credit quality without stabilizing intervention from rating agencies. Documented non-Markovian patterns can be attributed to the TTC perspective described as a four to six-year assessment of the default likelihood. Results do not support possible delays in rating revisions, but rather indicate small reactions in the magnitudes of the rating reviews. These conclusions are drawn from a new probability estimation technique focusing on hazard rates. It allows us to consistently estimate probabilities of default for all horizons at once, making all results robust. Our estimator is a smooth estimator of hazard rates which is free of bias and outperforms commonly used methodologies. It avoids the Markovian framework and takes care of censoring