Country credit-risk rating aggregation via the separation-deviation model†

^†This paper is dedicated to the memory of Peter Hammer.

Abstract

Country credit-risk ratings are evaluated independently by several agencies. A common method of aggregating the ratings into a single rating is by taking their averages (the averaging method). We show here that an approach that captures the relative ranking of the countries given by each agency leads to an improved aggregate rating with respect to several criteria. The approach we use – the separation-deviation model – was proposed by Hochbaum. We compare the separation-deviation model with the averaging method for aggregating country credit-risk ratings provided by three different agencies. We show that the aggregate rating obtained by the separation-deviation model has fewer rank reversals (discrepancies in the rank ordering of the countries) than the aggregate rating obtained by the averaging method. We further prove several properties of the separation-deviation model, including the property that the aggregate rating obtained by the separation-deviation model agrees with the majority of agencies or reviewers, regardless of the scale used.

^†This paper is dedicated to the memory of Peter Hammer.

Keywords:

• network flow
• aggregate ranking
• country credit-risk ratings
• group rating
• group decision-making

Acknowledgements

The research of the first author is supported in part by NSF award No. DMI-0620677 and CBET-0736232. The second author is supported in part by NSF and by Consejo Nacional de Ciencia y Tecnología (CONACyT), México, under grant No. 160619. The authors are grateful to an anonymous referee who helped them to improve the presentation of this paper.

Notes

^†This paper is dedicated to the memory of Peter Hammer.

The concept of uniform quadratic functions is defined in Section 4.

We are grateful to Hammer et al. 7 for permission to use this data.