Introducing Greater Rigor— Methodological Approach

Abstract

This chapter presents an approach for identifying whether biases exist. Some methodological improvements to the current credit rating literature are introduced, which bore in mind the criticisms of the existing empirical studies. The approach controls for: (1) a core set of theorized and quantifiable indicators of debt quality; (2) a complex qualitative aspect of the credit rating process—the behavior of CRAs in trying to balance rating timeliness and rating stability and (3) fixed effects such as time invariant political and institutional characteristics of sovereigns within an ordered response framework. This framework allows us to estimate a lower threshold below which debt quality changes can lead to a downgrade and an upper threshold above which debt quality changes can lead to an upgrade. We allow the data to determine whether the upper threshold systematically varies across different groups of sovereigns. This forms a new modeling approach for more rigorously assessing the existence of bias.