Rating calibration

Abstract

In this work we deal with the question of whether the evaluation of selected rating agencies is equivalent in some sense or not and whether it is possible to fi nd a relationship between assessments. The fact that rating agencies aff ect not only fi nancial market participants (by publication of companies or states ratings) is undeniable. On the one hand, these agencies are criticized for the rating changes, which have infl uence for example credit conditions for rated entity. On the other hand, ratings have a growing number of users for which ratings have become one of the few clues in today's complicated conditions powered by global fi nancial crisis. For this purpose we use the calibration problem techniques, because calibration fi nds a relationship between two measurements, in our case between two independent assessments of rating agencies. Due to the nature of the data we assume that the relationship can be described by linear or quadratic function. So we use estimates derived for the one-dimensional linear calibration model with quadratic calibration function. This all we illustrate by the real ratings. The situation is complicated by the ordinal type data of the examined variables. Among other things, we discuss relations between ratings coming from particular rating agencies and evolution of this relation over time.