Herding behaviour among credit rating agencies

Abstract

[EN] The aim of this paper is to identify the in uence degree among the main rating agencies and the other variables that aect rating changes for sub-sovereign entities in Germany, Austria, Belgium, France, Italy and Spain, using a total of 32 territorial entities between 1996 and 2012. Due to the shortage of European sub-sovereigns with more than 2 ratings, we estimated six binary probit regressions as a combination of 3 rating agencies two to two. We conclude that Fitch is the most in uential agency on the other two rating agencies, but Standard and Poor's is the leader. There are other relevant factors that explain the probability of change of rating but are less powerful with respect to the change of rating. Among the dierent specications, the signicant variables are almost always the same. Being a Spanish region, downgrades of country's rating, being a heavily indebted region or a weak budget performance, increase the probability of change of rating. Aging regions and rising unemployment or population growth rate also increase the likeli- hood of rating change. On the other hand, regions with high density, Level of Intergovernmental transfers to total revenue or Number of notches' changes by Moody's agency given for a specic region in a previous year, reduce the probability of change rating.