The effect of the number of states on the validity of credit ratings

TL;DR

This paper investigates how the number of rating states influences the reliability of credit ratings over time, using matrix estimation methods to assess the validity of rating processes.

Contribution

It provides empirical evidence that the ideal number of rating states varies over time and introduces a method to quantify the validity of rating processes.

Findings

The effect of the number of states on rating dynamics changes over time.

Matrix estimation methods can test the validity of rating processes.

Assuming a fixed number of states may lead to likelihood loss in modeling.

Abstract

We explicitly test if the reliability of credit ratings depends on the total number of admissible states. We analyse open access credit rating data and show that the effect of the number of states in the dynamical properties of ratings change with time, thus giving supportive evidence that the ideal number of admissible states changes with time. We use matrix estimation methods that explicitly assume the hypothesis needed for the process to be a valid rating process. By comparing with the likelihood maximization method of matrix estimation, we quantify the "likelihood-loss" of assuming that the process is a well grounded rating process.